Chapter-1
Basic Equations
1.1 State of stress at a point: At a point there are three mutually perpendicular planes i.e.
orthogonal planes. Three stresses components, one normal and two shears, act on each plane.
Therefore, nine stress components must be known at each point to define completely state of
stress at a point. Therefore six stress components in the Cartesian ( x, y, z ) co-ordinate system.
x xy xz
yx y yz
zx zy z
Therefore there are only six components of stress at a point, three normal stresses and three shear
stresses. Therefore stress at a point is specified as
x xy xz
xy y yz (1.1)
xz yz z
Similarly, six stress components in the cylindrical ( r , , z ) co-ordinate system.
r r rz
r z (1.2)
z z
rz
1.2 State of strain at a point: Similar to stress considerations, the nine components of strain
are reduced to six independent components, three linear strains corresponding to the change in
line elements and three shear strains corresponding to the change in right angles formed by line
elements. Thus the state of strain at a point of a body in the Cartesian ( x, y, z ) co-ordinate
system can be expressed in the matrix form as
x xy xz
xy y yz (1.3)
yz z
xz
Similarly, six strain components in the cylindrical ( r , , z ) co-ordinate system
r r rz
r z (1.4)
z z
rz
1.3 Strain Displacement relationship: The six strain components, three linear strain and
three shear strains, at a point of the body are related to the three displacements u, v, and w by the
following expressions in the Cartesian ( x, y, z ) co-ordinate system
u v w
Normal strain: x , y , z
x y z
(1.5)
u v v w u w
Shear strain: xy , yz , xz
y x z y z x
Strain displacement relationship for cylindrical ( r , , z ) co-ordinate system
u 1 v u w
Normal strain: r , , z
r r r z
(1.6)
1 u v u v 1 w u w
Shear strain: r , z , rz
r r r z r z r
1.4 Equilibrium Equations: Consideration of the variation of the state of stress from point
to point leads to the equilibrium equations in the Cartesian ( x, y, z ) co-ordinate system are given
by
x xy xz xy y yz xz yz z
X 0; Y 0 and Z 0
x y z x y z x y z
(1.7)
Where X , Y and Z are the components of body force such as gravitational, centrifugal, or other
inertia forces. The equilibrium equations for a body referred in cylindrical co-ordinates ( r , , z )
system
r 1 r r 1 z 2 r
rz r Pr 0 ; P 0
r r z r r r z r (1.8)
rz 1 z z rz
and Pz 0
r r z r
Where Pr , P and Pz are the components of body force such as gravitational, centrifugal, or
other inertia forces.
1.5 Strain compatibility equations:
It is clear from the strain displacement relationship that if the three displacement
components are given, then the strain components can be uniquely determined. If, on the other
hand, the six strain components are arbitrarily specified at a point, then the displacement
components cannot be uniquely determined. This is because the six strain components are related
to only three displacement components viz u, v and w . Hence if displacement components are to
be single valued and continuous, then there must exist certain interrelationship among the strain
components. These relations are called the strain compatibility equations. For three dimensional
bodies there exist six strain compatibility equations.
In the Cartesian ( x, y, z ) co-ordinate system.
2 x y xy 2 y 2 z 2 yz 2 x 2 z 2 xz
2 2
; and (1.9)
y 2 x 2 xy z 2 y 2 yz z 2 x 2 xz
2 x xy xz yz 2 y
xy xz yz
yz x z x xz y z x
2 ; 2
y y
And (1.10)
2 z xz yz xy
2
xy z y x z
Similarly strain compatibility equations, for the case of small displacements, in terms of
cylindrical coordinates ( r , , z ) can be obtained as
2 r 2 z 2 rz r 2 r 2 r 2 r r
; r r
z 2 r 2 rz r 2 r 2 r (1.11)
2
z
2
z
2
r 2 2 r z r rz r
z r 2
z z
r r r z rz 2r r r
z r z r
r 2 1 rz r r
2
r r
2
And r z 2 (1.12)
r r r r z z
r 1 rz 1 z
r z 2
z z r r r r r
1 1
r r z ; r z
E E
(1.14)
1 r z rz
z z r and r , z , rz
E G G G
Alternately stress-strain relation for isotropic material can be written as,
1 0 0 0
x 1 0 0 0 x
0
y 1 0 0
y
z 1 2 0 z
E
0 0 0 0 (1.15)
xy 1 1 2 2 xy
1 2 0
zy 0 0 0 0 yz
zx 2 xz
1 2
0 0 0 0 0
2
D
OR
x x y z 2 G x
y x y z 2 G y (1.16)
z x y z 2 G z
E E
Where Lame's constant = and G
1 1 2 2 1
r r z 2 G r
r z 2 G (1.17)
z r z 2 G z
1.7 Plane Stress Problems: The two dimensional elasticity problems under the following
conditions, is considered as plane stress problem.
1. The thickness of body i.e. the dimension in one direction, say z direction, is small in
comparison with other dimensions in x and y directions.
2. The load acting on the body are in the plane perpendicular to the thickness of the body
i.e. z axis. The loads are distributed uniformly over the thickness.
3. The stresses in the direction of thickness of the body must be zero on free boundary.
Therefore it is assumed that the stress components in the direction of thickness are zero
i.e. z xz yz 0
The plane stress problem is characterized by the following basic concepts and relations of theory
of elasticity.
1.7.1 State of a stress at a point: The three stress components in z direction are zero
x
y (1.18)
xy
1.7.2 State of a strain at a point: Strain at a point is defined by three independent strain
components, two linear strain, x and y , and one shear strain xy .
x
y (1.19)
xy
x y
It may be noted that z 0 , but can be obtained by the condition z 0 , z
E
1.7.3 Strain Displacement Relationship: Three strain components are related to two
displacement components by the following expressions:
u v u v
x ; y and xy (1.20)
x y y x
1.7.4 Equations of equilibrium: Conditions of equilibrium Fx 0, Fy 0 , for the
components of body force and internal forces due to stresses give the following equations.
x xy xy y
X 0 and Y 0 (1.21)
x y x y
1.7.5 Strain Compatibility Equations: The strain components x , y and xy are related by
eliminating displacements from the strain displacement relations as seen previously. Therefore
for plane stress problem, the strain compatibility equation is expressed as
2 x y xy
2 2
(1.22)
y 2 x 2 xy
1.7.6 Stress Strain Relationship: Stresses can be expressed in terms of strains as per Hooks
law
x 1 0 x
E
y 1 0 y (1.23)
1
2
1
xy 0 0 xy
2
1.8 Plane Strain Problems: The two dimensional elasticity problems under the following
conditions, is considered as plane strain problem.
1. The thickness of body i.e. the dimension in one direction, say z direction, is very large in
comparison with other dimensions in x and y directions.
2. External forces are perpendicular to the z axis.
3. The strains in the z direction of the length of the body are zero z xz yz 0
The plane strain problem is characterized by the following basic concepts and relations of theory
of elasticity.
1.8.1 State of a stress at a point: Stress at a point is defined by three independent stress
components, two normal stress, x and y , and one shear stress xy .
x
y But z 0 (1.24)
xy
It may be noted that z 0 , but can be obtained by the condition z 0 , z x y
1.8.2 State of a strain at a point: The three strain components in z direction are
zero z xz yz 0 . Therefore only three independent strain components
x , y and xy exist at a point. Note that xy yx . Such state of strain in known as plane strain.
x
y (1.25)
xy
1.8.3 Strain Displacement Relationship:
The displacement in z direction is zero. Therefore the two displacement components u, v in
x and y directions, exists at a point. The relation between strain components and displacement
components are same as before, i.e.
u v u v
x y and xy
y x
; (1.26)
x y
1.8.4 Equations of equilibrium: Corresponding to three stress components in xy plane, the
equilibrium equations are also same and restated as
x xy xy y
X 0 and Y 0 (1.27)
x y x y
1.8.5 Strain Compatibility Equations: eliminating displacements components, the three strain
components are related by one equation as proved already.
2 x y xy
2 2
(1.28)
y 2 x 2 xy
1.8.6 Stress Strain Relationship: These relations are different than that of plane stress problem
and can be derived from the strain condition z 0 . Stresses can be expressed in terms of strains
as per Hooks law
x 1 0 x
E
y 1 0 y (1.29)
1 1 2
xy 0 1 2
xy
0
2
Chapter 2
Bending of Rectangular Thin Plates
2.1 Introduction:
Flat plates are extensively used in many engineering application like roof and floor of
buildings, deck slab of bridges, foundation of footing, water tanks, bulk heads, turbine disks etc.
plates used in such applications are normally subjected to lateral loads, causing bending of the
plate. The geometry of the plate normally defined by the middle plane which is plane
equidistance from the top and bottom faces of the plate. The thickness of the plate (h) is
measured in a direction normal to the middle plane of the plate. The flexural properties of the
plate largely depend on its thickness rather than its two dimensions (length and width). Plate is
subjected to moments and forces as shown in figure, transverse shear forces acts perpendicular to
the plane of the plate whereas central shear forces acts in the plane of the plate. In general, plate
problems can be classified into three major categories viz. thin plate, moderately thick plate and
thick plate, depending upon the thickness of the plate.
If the thickness of the plate is very small as compared to the other two dimensions, then such a
plate can be called as thin plate. A plate can be considered as thin if the ratio of thickness to the
lesser of other two dimensions is less than 0.05. The simplest and most widely used plate theory
is classical small deflection theory of thin isotropic and anisotropic plates.
2.2 Small Deflection Theory / Kirchhoffs Thin Plate Bending Theory / Classical Plate
Theory:
The classical thin plate theory is based on Love-Kirchhoffs hypothesis which makes
assumptions similar to those made by the Bernoulli-Navier hypothesis used in the theory of thin
or shallow beams. If the deflection is not small, then the bending of the plate is accompanied by
in plane stresses. The nature and magnitude of these in plane stresses depends not only
magnitude of the deflection but also type of support provided at the edges of the plate.
2.2.1 Assumptions in small deflection theory:
The following fundamental assumptions are made in the classical small deflection theory of
thin homogenous elastic plates.
1. Straight line initially normal to the middle surface to the plate remains straight and normal to
the deformed middle surface of the plate and unchanged in length.
Note: This assumption corresponds to the Bernoullis melier hypothesis for the deflection of
beam.
2. Displacement w is assumed to be very small. This means the slope of the deflected surface is
small and hence square of the slope would be negligible in comparison with unity
3. The normal stresses x and y , inplane shear stress xy are assumed to be zero at middle
4. Stress z i.e. transverse normal stress is small as compared to other stress components and
may be neglected in stress strain relationship.
z x , y , xy
u
sin
z
w u
sin
x z
w w w
u z sin
x x x
Under the second assumption, points on the middle surface of the plate can be assumed to be
displaced only in the z direction. For other points of plate, the corresponding u and v
displacements in the x and y direction are proportional to their distances from the middle surface.
w
Thus point is displaced by the amount u z in the x direction and by the amount
x
w
v z in the y direction. Therefore displacements in the x, y and z directions are given as.
y
w
u z a
x
w
v z b (2.1)
y
w w x, y c
To obtain components of strain put equation (2.1) in the (1.20)
2w
x z a
x 2
2w
y z 2 b (2.2)
y
2w
xy 2 z c
xy
To find out corresponding stresses put (2.2) in the equation (1.23)
z E 2w 2w
x a
1 2 x 2 y 2
z E 2w 2w
y b
1 2 y 2 x 2
(2.3)
2w E 2w
xy 2G c
xy 1 xy
consider the variation of forces on each edge, but instead assume that the mean value of forces
would be acting at the centre of each face or edge. The bending and twisting moments and shear
forces acting on the plate area of unit width and height h are shown in following figure.
Where
M x , M y Bending moments on a plane normal to the x and y axes respectively
It is assumed that positive moments produce tension in the fibres located at the bottom part of
plate. Now, since moments are resultant of the stresses developed in the plate, these are called
stress resultant and are forces per unit length of the plate. Which are given by
h / 2 h / 2
Mx
h / 2
x .b .z .dz .dy
h / 2
x .z .dz (2.4)
Substitute the value of x from equation (2.3 a) in the equation (2.4) to obtain M x
h/ 2
2 w 2w
2
E
Mx
1 2 x 2
h / 2
z dz.
y 2
h/ 2
E 2w 2w z3
Mx
1 2 x 2 y 2 3 h / 2
Eh3 2w 2w
Mx 2
12 1 2 x y 2
2w 2w
Mx D 2 (2.5)
x y 2
Eh3
Where D Flexural Regidity of Plate and w denotes small deflection of the plate
12 1 2
Substitute the value of y from equation (2.3 b) into the equation (2.6) to obtain M y , hence
2w 2w
My D 2 (2.7)
y x 2
Now, the shearing stresses xy give the twisting moment. Therefore twisting moment will be
Substitute the value of xy from equation (2.3 c) into the equation (2.8) to obtain M xy , hence we
get
h/ 2
E 2w 2
M xy
1
h / 2
xy
z dz.
h/ 2
E 2w z3
1 xy 3 h / 2
M xy
M xy
Eh3 1 2 w
12 1 1 xy
Eh3 1 2w
M xy
12 1 2 xy
2w
M xy D 1 (2.9)
xy
Now taking moments of all forces on the element with respect to x axis and equating to zero, we
obtained the equation of equilibrium
Mx 0
M y M xy
My dy dx M y dx M xy dx dy M xy dy
y x
Qy dy Qx dy dy
Qy dy dx dy Qx dy Qx dx dy q x , y dx dy 0
y 2 x 2 2
M y M xy
dx dy dx dy Qy dx dy 0
y x
M y M xy
Qy 0 (2.10)
y x
While obtaining above equilibrium equation, the moment due to lateral load q and the moment
Qy
due to the rate of change of shear forces viz have been neglected since these quantities are
y
multiplied by terms which are very small.
M y M xy 2w 2w
Qy D 2 (2.11)
y x y x 2 y
In the same manner, by taking moments with respect to the y axis ( M y 0 ), we get
M x M xy
Qx 0 (2.12)
x y
M x M xy 2w 2w
Qx D 2 (2.13)
x y x x 2 y
Considering all the forces acting on the element in the z direction we obtain the following
equation of equilibrium. If q x, y = intensity of distributed external load on the plate surface, then
Qx Qy
Qx dx dy Qx dy Qy dy dx Qy dx q x ,y dx dy 0
x y
Qx Qy
dx dy dx dy q x ,y dx dy 0
x y
Qx Qy
q x , y 0 (2.14)
x y
Since there are no forces in the x and y directions and no moments with respect to the z axis, the
three equations (2.10), (2.12) and (2.14) completely define the equilibrium of the element. Let us
eliminate the shearing forces Qx and Qy from these equations by determining them from
equations (2.10) and (2.12) and substitute into the equation (2.14). Hence
M x M xy M y M xy
q x ,y 0
x x y y y x
2 M x M xy M y M xy
2 2 2
q x,y 0
x 2 xy y 2 xy
2M x 2 M xy 2 M y
2 q x,y 0 (2.15)
x 2 xy y 2
To represent this equation in terms of the deflections w of the plate, use equations (2.5), (2.7) and
(2.9), we obtain
2 2w 2 w 2 2w
D 2
xy
2 D 1
x 2 x y 2 xy
2 2w 2w
D 2 2
q x ,y 0
y 2 y x
4w 4w 4w 4w 4w
D 4 2 2 2 1 2 2 4 2 2 q x ,y 0
x x y x y y x y
4w 4w 4w q x,y
2 (2.16)
x 4
x y
2 2
y 4
D
This latter equation can also be written in the symbolic form
2
2 2 q x ,y
2 2 w
x y D
q x ,y
2 2
w
D
2
2 2
Where 2 2 Laplacian operator
2
x y
q x , y
4 w (2.17)
D
Which is the Lagrange equilibrium equation for the bending of thin plate with constant
thickness. Therefore the solution of the problem of bending of plates by a lateral load reduces to
the integration of equation (2.17). If, for a particular case, a solution of this equation is found that
satisfies the conditions at the boundaries of the plate, then bending and twisting moments can be
calculated from the equations (2.5), (2.7) and (2.9). The corresponding normal and shearing
stresses are obtained from equation (2.3).
Mx M h 6Mx
x Z 3 x
I
h /12 2 h2
6 M x max
x max (2.18)
h2
6My 6 M xy
Similarly y max 2
and xy max (2.19)
h h2
Equations (2.11) and (2.13) are used to determine the shearing forces Qx and Qy . The shearing
stresses xz and yz can now be determined by assuming that they are distributed across the
xz max
3 Vx
2 h
and
yz max
3 Vy
2 h
(2.20)
w y b 0
2w 2w (2.23)
My D 2 2 0
y b
y x y b
Since the changes of w w.r.t. x co-ordinate vanishes along the edge
2w
My D 2 0 (2.24)
y y b
ya
3) Free edge condition: In the most general case a twisting moment, bending moment and
transverse shear force acts on an edge of the plate. An edge on which all these three stresses
vanishes is defined as free edge.
M x M xy Qx 0
xa
(2.27)
M y M yx Qy 0
y b
Kirchhoff Paradox: Later on Kirchhoff provided that three boundary conditions are too
many and so that two conditions are sufficient for the complete determination of w satisfy the
q
equation 4 w . Kirchhoff pointed out that the two conditions prescribing M xy and Qx can
D
be replaced by a single one. The reason is that the twisting moment acting on an element of the
edge of the plate may be replaced by two statically equivalent vertical forces, which can then
combined with the vertical shearing forces.
Note: owing to such replacement the stress distribution in the intermediate neighborhood of the
edge will naturally be affected but the stress distribution in the rest of the plate essentially the
same.
The corner of rectangular plate under the action of uniformly distributed load tends to
rise. This action is prevented by the concentrated reactions at the corners.
2.3 Naviers Solution for Lateral Deflection of Simply Supported Rectangular Plate:
The solution of preceding article can be used in calculating deflections produced in a
simply supported rectangular plate y any kind of loading given by the equation q q x, y .
Consider a rectangular plate of sides a and b which is simply supported on four sides and
subjected to distributed load q , which is function of two variables x and y can be expressed in
the domain 0 x a , 0 y b by a double trigonometric series given by
m x n y
w w
m 1 n 1
mn sin
a
sin
b
(2.29)
m x n y
q x,y q
m 1 n 1
mn sin
a
sin
b
where m,n 1,3,5......... (2.30)
To calculate any particular coefficient qmn of this series, we multiply both sides of equation
(2.30) by sin j x / a sin k y / b and integrate with respective x and y from 0 to a and 0 to
b. observing that
x a y b x a y b
j x k y m x n y j x k y
q x,y sin sin dx dy qmn sin sin sin sin dx dy (2.31)
x 0 y 0
a b x 0 y 0
a b a b
x a
m x j x
Performing integration with respect to x, we get sin
x 0
a
sin
a
dx 0 at m j
yb
n y k y
Performing integration with respect to y, we get sin
y0
b
sin
b
dy 0 at n k
xa
1 m x
2 x0
1 cos 2
dx
a
xa
1 m x a
x sin 2 a 2m
2 x0
x a
m x j x a
sin
x 0
a
sin
a
dx
2
(2.32)
x a y b
4 m x n y
qmn
ab q
x 0 y 0
x, y
sin
a
sin
b
dx dy (2.34)
Performing the integration indicated in equation (2.34) for a given load distribution, i.e., for a
given q x,y , we find the coefficient of series (2.30) and represent in this way the given load as a
sum of partial sinusoidal loadings. Now, Substitute the solution of w (2.29) and q x,y (2.30) in
qmn
wmn (2.35)
m m2 n2 4 n4 4
4 4
D 4 2 2 2 4
a a b b
Substitute equation (2.35) in the equation (2.29)
qmn m x n y
w
m 1 n 1 m4 4 m2 n 2 4 n 4 4
sin
a
sin
b
(2.36)
D 4 2 2 2 4
a a b b
4 2a 2b
qmn m n
ab
16 q0
for m, n 1, 3, 5 ..........
qmn mn 2 (2.39)
0 for m, n 2, 4, 6 ..........
Substitute magnitude of wmn from equation (2.40) into the equation (2.36) to obtain the
deflection of the plate, therefore the expression for the deflected shape of the plate can now be
written as
16q0 1 m x n y
w mn 6
sin
2 sin
b
(2.41)
m 1 n 1 m 2
n 2 a
D 2 2
a b
Equation of Deflection for rectangular plate subjected to uniformly distributed load
In case of uniform load we have a deflection surface symmetrical with respect to axes
x a / 2, y b / 2; and quit naturally all terms with even numbers for m or n series (2.41)
vanish, since they are unsymmetrical with respect to the above mentioned axes. The maximum
deflection of plate is at its center and is found by substituting x a / 2, y b / 2; in the equation
(2.41).
m n
16q0 1
1
1
w 2 (2.42)
m 1 n 1 mn 6 m2 n2
2
D 2 2
a b
This is a rapidly converging series, and a satisfactory approximation is obtained by taking only
the first term of the series, which, for example in the case of square plate gives
4 q0 a 4 0.0416 q0 a 4
wmax (2.43)
6 D D
The expressions for bending moments and twisting moments can be obtained from equation
(2.5), (2.7) and (2.9). Since the moments are expressed by second derivatives of series,
differentiate equation (2.41) w.r.t. x and y.
m x n y
sin sin
w a b m
2 2 2
16q0
x 2
mn 6 2 2 (2.44)
m 1 n 1 m2 n2 a
D 2 2
a b
m x n y
cos sin
w a b m
3 3 3
16q0
x3
mn 6 2 3 (2.45)
m 1 n 1 m2 n2 a
D 2 2
a b
m x n y
sin sin
w a b n
2 2 2
16q0
y 2
mn 6 2 2 (2.46)
m 1 n 1 m2 n2 b
D 2 2
a b
m x n y
cos cos
w b mn
2 2
16q0 a
xy
mn 6 2 (2.47)
m 1 n 1 m2 n2 ab
D 2 2
a b
m x n y
cos sin
w b mn
3 2 3
16q0 a
xy 2
mn 6 2 2 (2.48)
m 1 n 1 m2 n2 ab
D 2 2
a b
Substitute derivatives from equations (2.44) and (2.46) into the equation (2.5) to obtain bending
moment in x direction i.e. M x
16q0 1 m2 2 n2 2 m x n y
M x D 2 2 sin sin
m 1 n 1 mn 6 m2
2
n2 a b a b
D 2 2
a b
16q0 1 m2 n2 m x n y
Mx mn m
4 2 2
2 sin sin
a b
(2.49)
m 1 n 1
2
n a
2 b
a 2 b2
Similarly Substitute derivatives from equations (2.44) and (2.46) into the equation (2.7) to obtain
bending moment in y direction i.e. M y
16q0 1 n2 m2 m x n y
My mn 4 m2 2 2
sin
a2
sin
a b
(2.50)
m 1 n 1 n2 b
a 2 b2
It is seen that maximum bending moments accurse at the centre of plate. Substituting x a / 2
16q0
1 m x n y
M xy 1
ab 4
n 1 m
2
cos
a
cos
b
(2.53)
m 1
2
n2
a 2 b2
The expression for vertical shear forces can be obtained from equation (2.11) and (2.13)
M xy M xy
Qx Vx And Qy Vy (2.54)
y x
In which Vx and Vy are the reactive forces at the supported edges of the plate. Therefore the
expressions for resultant shears or reactive forces at the supported edges of the plate can be
obtained from following equations. For the edges x = a we find
M xy 3w 3w
Vx Qx x3
xy 2
D 2 (2.55)
y x a
In the same manner reactive force for the edge y = b,
M xy 3w 3w
Vy Qy y 3
x 2y
D 2 (2.56)
x y b
Therefore reactive force in x direction is given by substituting equation (2.45) and (2.48) into the
equation (2.55), we obtain
16q0 1 m x n y m
3 3
cos sin
m 1 n 1 mn D m n a b a
6 2 3
2 2
a2 2
b
Vx D
2
m x n y mn
2 3
16q0 1
cos sin
m 1 n 1 mn
2 2
a b ab
6 2
m 2
n
D 2 2
a b
16q0 1 m3 mn2 m x n y
Vx
m 1 n 1 mn m
3 2 3
2 2 cos sin
a b
(2.57)
2
n2 a ab
a 2 b2
16q0 1 m3 mn2 n y
Vx x a
m 1 n 1 mn m
3 2 3
2 2 sin
b
(2.58)
2
n2 a ab
a 2 b2
Similarly, reactive force in y direction is given by using equation (2.56)
16q0 1 n3 m2 n m x n y
Vy
m 1 n 1 mn m
3 2 3
2 2 sin
a
cos
b
(2.59)
2
n2 b ab
a 2 b2
n3 m2 n m x
Vy
16q0 1
yb
m 1 n 1 mn m
3 2 3
2 2 sin
a
(2.60)
2
n2 b ab
a 2 b2
The minus sign indicates that the reactions on the plate act upward. From symmetry it may be
concluded that equations (2.58) and (2.60) also represents pressure distribution along the
sides x 0 and y 0 , respectively.
From equation (2.28) the concentrated reaction at each corner can be determined as
32q0
1 m x n y
R 1
ab 4
n 1 m
cos
a
2 cos
b
(2.61)
m 1
2
n 2
a 2 b2
Total downward force = 4R
128q0
1 m x n y
4 R 1
ab 4
n 1 m
cos
a
2 cos
b
(2.62)
m 1
2
n2
a 2 b2
Now bending stresses in x and y directions are determined using equation (2.18), therefore
bending stress in x direction is given by
96q0 1 m2 n2 m x n y
x h2 mn 4 m2 2 2
2 2
sin sin
a b
(2.63)
m 1 n 1 n a b
a 2 b2
The maximum bending stress is at the center of the plate i.e. x a / 2 and y b / 2 , hence the
maximum bending stress in x direction is
96 q0 1 m2 n2
x max h2 mn 4 m2 2 2
b2
(2.64)
m 1 n 1 n2 a
a 2 b2
Similarly, Bending stress in y direction
96 q0 1 n2 m2 m x n y
y h mn m
2 4 2 2
2
sin
sin
b
(2.65)
m 1 n 1
2
n 2
b a a
a 2 b2
The maximum bending stress is at the center of the plate i.e. x a / 2 and y b / 2 , hence the
maximum bending stress in x direction is
n2 m2
y 96 q0 1
h2 mn 4 m2 2 2
a2
(2.66)
n2 b
max
m 1 n 1
a 2 b2
The maximum shearing stress will be at the middle of the sides of the plate. Observing that the
total transverse force Vx and Vy is distributed along the thickness of plate according to parabolic
a2
b2
2.3.2 Simply supported Plate carrying Sinusoidal load:
Taking coordinate axes as shown in figure, we assume that the load distributed over the surface
of the plate is given by the expression
x y
q x, y q0 sin sin (2.69)
a b
In which q0 represents the intensity of load at the centre of plate as shown in figure. Substituting
equation (2.69) in the equation (2.34) and after integration we get Fourier coefficient
q for m 1 n 1
qmn 0 (2.70)
0 for m 1 n 1
It may seen from the boundary conditions are satisfied if we take for deflection the expression,
x y
w w1 sin sin (2.72)
a b
Substitute qmn q0 in the equation (2.35) we get
q0
w1 2
(2.73)
1 1
D 2 2
4
a b
Substitute equation (2.73) in the equation (2.36) and we conclude that the deflection surface
satisfying governing differential equation and boundary conditions.
q0 x y
w sin sin (2.74)
a b
2
4 1 1
D 2 2
a b
The expressions for bending moments and twisting moments can be obtained from equation
(2.5), (2.7) and (2.9). Differentiate equation (2.74) w.r.t. x and y we get
q0 1 1 x y
Mx 2 2 sin sin (2.75)
b
2
1 1 a b a
2 2
2
a b
q0 1 1 x y
My 2 2 sin sin (2.76)
1 b a a b
2
2 1
2 2
a b
q0 x y
M xy 1 cos cos (2.77)
a b
2
1 1
2 2 2
a b
It is seen that the maximum deflection and the maximum bending moments are at the centre of
plate. Substituting x a / 2 and y b / 2 in the equations (2.75), (2.76) and (2.77), we obtain
q0
w max 2
(2.78)
1 1
D 2 2
4
a b
q0 1
M x max 2 2 (2.79)
1 a b
2
1
2 2 2
a b
1
M y max
q0
2 2 (2.80)
1 b a
2
1
2 2 2
a b
In the particular cases of a square plate, a = b, and the forgiving formulae becomes
q0 a 4 1 q0 a 2
wmax
4D 4
M x max M
y max
4 2
(2.81)
We use equations (2.11) and (2.13) to calculate the shearing forces and obtained
q0 x y
Qx cos sin (2.82)
a b
2
1 1
a 2 2
a b
q0 x y
Qy cos sin (2.83)
a b
2
1 1
b 2 2
a b
Therefore the expressions for resultant shears or reactive forces at the supported edges of the
plate can be obtained from the equations (2.11) and (2.13).
M xy q0 1 1 x y
Vx Qx 3 2 2 cos sin (2.84)
y 1 1 a
2
ab a b
2 2
a b
For the edges x = a we find
q0 1 2 y
Vx x a 3 sin (2.85)
b
2 2
1 1 a ab
2 2
a b
Similarly,
M xy q0 1 1 x y
Vy Qy 3 2 2 sin cos (2.86)
x 1 1 b
2
a b a b
2 2
a b
For the edges y = b we find
1 2 x
V
y yb
q0
3 2 sin
2
a
(2.87)
1 1 b
2 2
ab
a b
Hence the pressure distribution follows the sinusoidal law. The minus sign indicates that the
reactions on the plate act upward. From symmetry it may be concluded that equations (2.85) and
(2.87) also represents pressure distribution along the sides x 0 and y 0 , respectively.
Now bending stresses in x and y directions are determined
6 q0 1 1 x y
x 2 2 2 sin sin (2.88)
h 1 a b a b
2
2 1
2 2
a b
6 q0 1 1 x y
y 2 2 2 sin sin (2.89)
h b
2
1 1 b a a
2 2
2
a b
The maximum bending stress is at the centre of the plate. Hence substitute x a / 2 and y b / 2
To fine the maximum bending stress.
6 q0 1
x max 2 2 2 (2.90)
h 1 a b
2
1
2 2 2
a b
6 q01
y 2 2 2 (2.91)
h 1 b a
2
1
2 2 2
a b
The centre of the patch load will be at a distance x0 from the origin in the x direction and y0
from the origin in the y direction. Patch load is acting over the area u v , therefore the intensity
The Fourier coefficient qmn for this case can be determined from equation (2.34). Therefore
x0 u / 2 y0 v / 2
4P m x n y
qmn
a bu v
x0 u / 2
y0 v / 2
sin
a
sin
b
dx dy (2.92)
a m x0 m u m x0 m u
cos a 2a cos a 2a
m
a m x0 m u
2 sin a sin 2a
m
x0 u / 2
m x 2a m x0 m u
x0 u / 2
sin
a
dx sin
m a
sin
2a
(2.93)
Therefore substitute equations (2.93) and (2.94) into the equation (2.34) to determined value of
coefficient qmn
4 P 2a m x0 m u 2b n y0 n v
qmn sin sin sin sin
a bu v m a 2a n b 2b
Hence the Fourier coefficient qmn for a patch load can be obtained as
16 P m x0 n y0 m u n v
qmn 2
sin sin sin sin (2.95)
mnu v a b 2a 2b
If patch load is acting at the centre of the plate, i.e. x0 a / 2 and y0 b / 2 , put into the equation
(2.95), we obtain
16 P m u n v
qmn sin sin (2.96)
mnu v 2a 2b
2
16 P
If u a and v b , equation (2.96) becomes qmn which is similar to Fourier
mn 2 a b
coefficient obtained for uniformly distributed load. Substitute the value of qmn in the above
Substitute magnitude of wmn from equation (2.97) into the equation (2.36) to determined
deflection w of the plate
m u n v
sin sin
16 P 2a 2b m x n y
w
m 1 n 1 mn u v 2
m
4 4
mn2 2 4
n
4 4
sin
a
sin
b
(2.98)
D 4 2 2 2 4
a a b b
The series converge rapidly, and we can obtain the deflection at any point on the plate with
sufficient accuracy by taking only first few terms of the series. Let us, for example, calculate the
maximum deflection, i.e. when patch load at the center of the plate i.e. x a / 2 and y b / 2
16 P 1 m u n v
wmax sin sin (2.99)
m 1 n 1 mn u v 6 m 4
mn2 2
n
4
2a 2b
D 4 2 2 2 4
a a b b
Fourier coefficient for the plate carrying concentrated load can be determined from equation
(2.95) by making some mathematical adjustment. Hence using following mathematical
adjustment in the equation (2.95), we obtain
m u n v
m x sin sin
16 P n y0 2a m u 2b n v
qmn sin 0
sin
mnu v 2 a b m u 2a n v 2b
2a 2b
m u n v
sin sin
4P m x0 n y0 2a 2b
qmn sin sin
ab a b m u n v
2a 2b
Hence after simplification the Fourier coefficient qmn for a concentrated load can be obtained as
4P m x0 n y0 sin
qmn sin a sin b lim 1 (2.100)
ab 0
Substitute the value of Fourier coefficient qmn in the above equation (2.35) to determine value of
4P 1 m x0 n y0
wmn
m 1 n 1 ab m4 4 m2 n 2 4 n 4 4
sin
a
sin
b
(2.101)
D 4 2 2 2 4
a a b b
Substitute magnitude of wmn from equation (2.101) into the equation (2.31) to determined
deflection w of the plate.
m x0 n y0
sin sin
4P a b m x n y
w
m 1 n 1 ab m 4 4
mn2 2 4
n
4 4
sin
a
sin
b
(2.102)
D 4 2 2 2 4
a a b b
If concentrated load is acting at the centre of the plate, put x0 a / 2 and y0 b / 2 in the above
equation, we get
4P 1 m x n y
w
m 1 n 1 ab m 4
mn
2 2 44
n
4 4
sin
a
sin
b
(2.103)
D 4 2 2 2 4
a a b b
This is a rapidly converging series, and a satisfactory approximation is obtained by taking only
the first term of the series, which, for example in the case of square plate gives
P a2 0.0112 P a 2
wmax 4 (2.105)
D D
2.3.5 Simply supported plate carrying linearly varying load (hydrostatic pressure):
In this case let assumed simply supported rectangular plate subjected to linearly varying
load as shown in figure. Intensity of loading is zero at x 0 and maximum at x a i.e. q0 .
But for linearly varying load the value of q x, y at any point at a distance of x is given by the
q0 x
expression q x, y
a
The Fourier coefficient qmn for this case can be determined by substituting value of q x, y into the
equation (2.34)
x a yb
4q m x n y
qmn 20
a b
x 0 y0
x sin
a
sin
b
dx dy (2.106)
After performing integration with respect to x and y in the equation (2.106), we get
x a yb
m x a2 n y 2b
x 0
x sin
a
dx
m
And
y0
sin
b
dy
n
(2.107)
Substitute values of integration from equation (2.107) into the equation (2.106) to obtain the
Fourier coefficient qmn .
4q0 a 2 2b
qmn
a2 b m n
8q0
qmn (2.108)
mn 2
Substitute magnitude of Fourier coefficient qmn from equation (2.108) in the equation (2.35) to
Put magnitude of wmn into the equation (2.36) to obtained expression for deflection of the plate.
8q0 1 m x n y
w
m 1 n 1 mn 2 m mn
4 4
2 2 4
n
4 4
sin
a
sin
b
D 4 2 2 2 4
a a b b
8q0 1 m x n y
w Dmn m
6 2
sin
a
sin
b
(2.110)
m 1 n 1
2
n
2
a 2 b2
Deflection is maximum at the center of the plate, therefore substitute x a / 2 and y b / 2 in the
above equation, hence we get
8q0 1
wmax (2.111)
m 1 n 1 Dmn m2
6
n2
2
a 2 b2
This is a rapidly converging series, and a satisfactory approximation is obtained by taking only
the first term of the series, which, for example in the case of square plate gives
2q0 a 4
wmax (2.112)
D 6
In figure a plate a b bounded by x-y co-ordinates. The origin is o is taken at the midpoint of the
side b. the boundary condition at x 0 and x a are simply supported whereas those at
y b / 2 May arbitrary. A typical plate strip of unit width AB spanning along x is assumed to
behave as a simply supported beam of span a with the difference that the flexural rigidity of
this strip is D not EI. The applied load q q x, y is considered as prismatic i.e. along the x
w w1 x w2 x, y (2.113)
Solution for w1 x
For isotropic plate the particular solution is function of x only. Since it is solution of any
arbitrary plate strip AB as a beam satisfying the differential equation (2.16)
Since w1 x is a function of x only derivative with respective y vanishes from the equation.
4 w1 x q1 x
(2.114)
x 4 D
Where D is the flexural rigidity of the plate
It is not essential that particular solution has to satisfy all the boundary conditions at the four
edges. Here it satisfy all the boundary conditions at the two edges viz at x = a and x = a both of
which simply supported, i.e.,
w1 0 at x 0 and x a
2 w1 (2.115)
0 at x 0 and x a
x 2
The expression for w1 is obtained from equation (2.114). Let the solution of w1 and q1 can be
expressed interms of sine series
m x
w1 w
m 1
m sin
a
(2.116)
m x
And q1 q
m 1
m sin
a
(2.117)
Substitute equation (2.116) and (2.117) in the equation (2.114) to obtained the constant wm
m x m
m x
4 4
1
m 1
wm sin
a a
4
D
q
m 1
m sin
a
qm a 4
wm
m 1 m4 4 D
(2.118)
To determine the value of qm , multiply equation (2.117) by sin j x / a both sides we get
xa xa
j x j x m x
x0 q1 sin a dx
x 0
qm sin
a
sin
a
dx (2.119)
xa
j x m x
Performing integration at j m , we get
x 0
sin
a
sin
a
dx 0
Put magnitude of above integration from equation (2.120) into the equation (2.119) to obtain the
value of Fourier coefficient qm . Hence
xa
m x qm a
x 0
q1 sin
a
dx
2
xa
2 m x
qm
a
x 0
q1 sin
a
dx (2.121)
4q0
qm (2.122)
m
Substitute (2.122) into the equation (2.118) we get
4q0 a 4
wm
m 1 m5 5 D
(2.123)
4q0 a 4 1 m x
w1
D m 1 m
5 5
sin
a
(2.124)
Let assume the solution of w2 x , y which is product of two functions, one is function of y i.e.
4 w2 m 4 4
m x
x 4
a4
Y m sin
a
m 1
w2 m
m x
4 2 2
x 2 y 2
a 2 m 1
Y m
II
sin
a
(2.126)
4 w2
m x
Ym sin
IV
y 4 m 1 a
IV m 2 2 II m 4 4 m x q0
m
m 1
Y 2 2
Ym 4
Ym sin
a
(2.127)
a a D
The general solution for the above fourth order differential equation is taken in the form as
y yc P.I .
m y m y m y m y
Am cosh Bm sinh C m sinh
qa 4 a a a a
Ym (2.128)
D m y m y
Dm cosh
a a
Since the loading is considered prismatic along y, identical boundary conditions at y b / 2
renders the problem symmetric about x axis. In that case the constant Cm and Dm will vanish
i.e. Cm Dm 0 in the equation (2.83) and for antisymmetric problems in which loading on
the plate maintains antisymmetry about x- axis even if the boundary conditions at y b / 2 are
Therefore
For the symmetric problems:
qa 4 m y m y m y
Ym Am cosh a Bm a sinh a (2.129)
D
And for the antisymmetric problems:
qa 4 m y m y m y
Ym Cm sinh a Dm a cosh a (2.130)
D
Therefore solution of symmetric problems is given by substituting equation (2.129) in the
equation (2.125) we get
qa 4 m y m y m y m x
w2 x, y Am cosh Bm sinh sin (2.131)
m 1 D a a a a
Therefore to find expression for deflection of surface substitute equation (2.124) and (2.131) in
the equation (2.113) we get
4q0 a 4 m x
qa 4 m y m y m y m x
w
m 1 m D
5 5
sin
a
m1 D Am cosh
a
Bm
a
sinh sin
a a
qa 4
4 m y m y m y m x
w
D
m 1
m5 5 Am cosh a Bm a sinh a sin a
(2.132)
Above equation satisfy the boundary condition and equation of equilibrium of the plate.
Now to determine the constants in the equation (2.132) Am and Bm use boundary conditions
given below
w0 for y b / 2 (2.133)
2w
0 for y b / 2 (2.134)
y 2
Therefore using boundary condition from equation (2.133), put y b / 2 into the equation
(2.132) and equate with zero, we get
qa 4
4 m y m y m y m x
D
m 1
m5 5 Am cosh a Bm a sinh a sin a 0
4 m b m b m b
Am cosh Bm sinh 0
m
5 5
2a 2a 2a
m b
Let m
2a
4
Am cosh m Bm m sinh m 0 (2.135)
m 5
5
m y m m y m y m
2 2 2 2
Am cosh a a 2 Bm a sinh a a 2
w sin m x
2
qa 4
y 2
D
m 1
m y m
2 2
m y m
2 2 a
(2.136)
Bm cosh Bm cosh
a a a a
2 2
Substitute y b / 2 into the equation (2.136) and use boundary condition from equation (2.134)
and equate with zero, hence
Am cosh m Bm m sinh m 2 Bm cosh m 0 (2.137)
Subtract equation (2.137) from (2.135) we get
4
Am cosh m Bm m sinh m Am cosh m Bm m sinh m 2 Bm cosh m 0
m 5
5
4
2 Bm cosh m
m 5
5
2
Bm (2.138)
m cosh m
5 5
Substitute value of Bm from equation (2.138) into the equation (2.135) to obtained constant Am
4 2
Am cosh m m sinh m 0
m 5 5
m cosh m
5 5
2 m tanh m 4
Am cosh m 5 5
m 5 5
m
2 tanh 4
Am cosh m m 5 5 m
m
2 tanh m 4
Am 5m 5 (2.139)
m cosh m
Substitute constants Am and Bm from equations (2.138) and (2.139) into the equation (2.132) to
calculate the deflection of the plate.
4 2 m tanh m 4 m y
5 5 5 5 cosh
m m cosh m a m x
4
qa
w m y a
sin
D m 1 2 m y
5 5 sinh
m cosh m a a
m tanh m 2 m y
1 5 5 cosh
4qa 2m cosh m
4
a m x
w m5 5 D
sin
m 1 2 m y m y a
5 5 sinh
m cosh m a a
m tanh m 2 m y
5 5 cosh
4qa 4qa 2m cosh m a m x
4 4
w m D
5 5
m D
5 5
m y
sin
m y a
m 1
2
m5 5 cosh sinh
m a a
m tanh m 2 m y
5 5 cosh
4
5qa
4qa 4 2m cosh m a m x
w
384 D
m 1 m D
sin (2.140)
m y m y a
5 5
2
m5 5 cosh sinh
a
m a
This series in this expression rapidly converge and sufficient accuracy is obtained by taking only
the first term. Differentiate above equation with respective x and y to determine bending
moments and bending stresses.
Differentiate equation (2.140) with respect to x upto second order, we get
m tanh m 2 m y
5 5 cosh
2w m2 2 5qa 4
4qa 4 2m cosh m a m x
x 2
2
a 384 D
m 1 m D
sin (2.141)
m y m y a
5 5
2
m5 5 cosh sinh
a
m a
Similarly differentiate equation (2.140) with respect to y upto second order, we get
m tanh m 2 m y m
5 5 sinh
2m cosh m a a
w 5qa 4
4qa 4 2 m y m y m m x
y 384 D
5 5
m D m cosh m a
cosh sin (2.142)
a a a
5 5
m 1
2 m m y
5 5 sinh
m cosh m a a
m tanh m 2 m y m
2 2
5 5 cosh 2
2 m cosh m a a
w 5qa 4
2 4
4qa 2 m y 2 2
m y m m x
y 2 384 D
m 1
5 5
m D m cosh m a
5 5
sinh 2 sin
a a a
(2.143)
m y m
2 2
5 5
4
cosh
2
m cosh m a a
Bending and twisting moments can be calculated using equations (2.5), (2.7) and (2.9) and
corresponding normal stresses are found from equation (2.18) and (2.19)
2.4.2 Levys solution for rectangular plate with at least two opposite edges simply
supported carrying linearly varying load:
In this case also the solution w w x, y of the plate can be taken as sum of complimentary
Solution for w1 x
Let assume the solution scheme for w and q from equation (2.116) and (2.117). In case of
q0 x
uniformly varying load, value of load q1 at any point at a distance of x is given by q1
a
Substitute value of q1 into the equation (2.121) to find out the value of Fourier coefficient qm .
xa
2 m x
qm
a
x 0
q1 sin
a
dx
xa
2 q0 x m x
qm
a
x 0
a
sin
a
dx
xa
2q m x
qm 20
a
x 0
x sin
a
dx
2q m x a
a
m x a
a
qm 20 x cos sin
a a m 0 a m 0
2q0
qm (2.144)
m
Therefore substitute value of Fourier coefficient in the equation (2.118) to obtain constant wm
2q0 a 4
wm
m 1 m5 5 D
(2.145)
Put equation (2.145) into the equation (2.116) to obtained deflection of the plate w1 ( x)
2q0 a 4 m x
w1 x
m 1
sin
m D a
5 5 (2.146)
This should satisfy the plate equation and boundary condition at the edges y b / 2
Prof. Atteshamuddin S. Sayyad 50
Theory of Plates and Shells
Solution for w2 x, y
qa 4
2 m y m y m y m x
w
D
m 1
m5 5 Am cosh a Bm a sinh a sin a
(2.147)
Above equation (2.147) will satisfy the boundary condition and equation of equilibrium.
To determine the constant Am and Bm use boundary conditions from equation (2.133) and
(2.134).
Therefore using boundary condition from equation (2.133) substitutes y b / 2 into the equation
(2.147) and equate with zero, hence we get
2 m b m b m b
Am cosh Bm sinh 0
m5 5
2a 2a 2a
m b
Let m
2a
2
Am cosh m Bm m sinh m 0 (2.148)
m 5
5
(2.149)
m y m m y m y m
2 2 2 2
Am cosh a a 2 Bm a sinh a a 2
w sin m x
2
qa 4
y 2
D
m 1
m y m
2 2
m y m
2 2 a
(2.150)
Bm cosh Bm cosh
a a a a
2 2
Using boundary condition from equation (2.134) substitutes y b / 2 into the equation (2.150)
and equate with zero, hence we get
Am cosh m Bm m sinh m 2 Bm cosh m 0 (2.151)
Subtract equation (2.151) from (2.148)
2
Am cosh m Bm m sinh m Am cosh m Bm m sinh m 2 Bm cosh m 0
m 5
5
2
2 Bm cosh m
m 55
1
Bm (2.152)
m cosh m
5 5
Substitute value of constant Bm from equation (2.152) into the equation (2.148) to obtain Am
2 1
Am cosh m m sinh m 0
m 5 5
m cosh m
5 5
1 2
Am cosh m m sinh m 5 5
m cosh m
5 5
m
m tanh m 2
Am cosh m 5 5
m5 5
m
tanh 2
Am cosh m m 5 5 m
m
tanh m 2
Am m5 5 (2.153)
m cosh m
Substitute values of constants Am and Bm from equations (2.153) and (2.152) into the equation
(2.147) to obtained deflection of the plate
m y
2 sinh
qa4
tanh m 2 m y a m y sin m x
w
D
m 1
5 5 m5 5
m m cosh m
cosh
a m cosh m a a
5 5
(2.154)
Differentiate above equation (2.154) with respect to x and y to determine bending and twisting
moments from equations (2.5), (2.7) and (2.9), and corresponding stresses from equations (2.18),
(2.19) and (2.20)
2.4.3 Levys solution for rectangular plate carrying moments along edges:
Consider a rectangular plate simply supported all along the edges is acted on by moment
M distributed along the edges at y b / 2 . The solution can be efficiently carried out using
levys method. Since the acting load is the moment and not distributed loading the particular
integral part w1 x is absent. Let assume the solution of w2 x , y from equation (2.125)
m x
w w2 x, y Y
m 1
m sin
a
If the moment acting at the edges y b / 2 are M1 and M2 resp. and are an unequal in
magnitude.
Note: since acting load is moment all the four constants ( Am , Bm , Cm and Dm ) will be
2w
at y b / 2 w 0; M y D 2 M 2 (2.156)
y
If M1 M2 M then solution is given by equation (2.84) (symmetrical problem) and
(2.129)
m y m y m y m x
w
m 1
Am cosh a Bm a sinh a sin a
(2.157)
m
Let assume m
a
w
m 1
Am cosh m y Bm m y sinh m y sin m x (2.158)
w0 for y b / 2 (2.159)
2w
D 2 M for y b / 2 (2.160)
y
Using boundary condition (2.159), Substitute y b / 2 in the equation (2.158) and equate with
zero, we obtain
b b b
Am cosh m Bm m sinh m 0
2 2 2
b
Let assume m m , therefore equation becomes
2
Am cosh m Bm m sinh m 0 (2.161)
Now differentiate equation (2.158) with respective y upto second order
w
y
m 1
Am sinh m y m Bm m y cosh m y m Bm sinh m y m sin m x
2w
y 2
m 1
Am cosh m y m2 Bm m y sinh m y m2 2Bm cosh m y m2 sin m x (2.162)
Using boundary condition (2.160), Substitute y b / 2 in the equation (2.162) and equate
with M , hence we get
m b 2 m b m b 2 m b 2 M
Am cosh 2 m Bm 2 sinh 2 m 2Bm cosh 2 m sin m x D
4 M0
Where M is uniformly distributed over the edges and therefore its value is given by M ,
m
Substitute in the above equation we get
m b 2 m b m b 2
Am cosh 2 m Bm 2 sinh 2 m
sin m x 4 M 0 sin m x
m b 2 Dm
2 Bm cosh m
2
4M 0
Am cosh m Bm m sinh m 2 Bm cosh m (2.163)
m m2 D
Subtracting equation (2.163) from (2.160), we get
4M 0
Am cosh m Bm m sinh m Am cosh m Bm m sinh m 2 Bm cosh m
m m2 D
4M 0
2 Bm cosh m
m m2 D
2 M0
Bm (2.164)
m D cosh m
2
m
Substitute magnitude of constant Bm from equation (2.164) into the equation (2.161) to find
2 M0
Am cosh m m sinh m 0
m D cosh m
2
m
2 M 0 m tanh m
Am (2.165)
m m2 D cosh m
Therefore substitute values of constants Am and Bm from equations (2.165) and (2.164) into the
equation (2.158) to find out deflection of the plate.
2 M 0 m tanh m 2 M0
w cosh m y m y sinh m y sin m x
m m D cosh m m m D cosh m
2 2
m 1
2 M0
w m
m 1
2
m tanh m cosh m y m y sinh m y sin m x
D cosh m
(2.166)
m
2.4.4 Levys solution for rectangular plate with two opposite edges Clamped subjected to
uniformly distributed load:
When two opposite edges of the plate are clamped and subjected to uniformly distributed load,
the deflection of the plate is given by the equation (2.132)
qa 4
4 m y m y m y m x
w
D
m 1
m5 5 Am cosh a Bm a sinh a sin a
Applying the boundary conditions at clamed edges to find out unknown constants Am and Bm
w0 for y b / 2 (2.167)
w
0 for y b / 2 (2.168)
y
Therefore substitute y b / 2 in the above equation of deflection and equate with zero using
boundary condition (2.167)
4 m b m b m b
Am cosh Bm sinh 0
m5 5
2a 2a 2a
m b
Let m
2a
4
Am cosh m Bm m sinh m 0 (2.169)
m 5
5
w
Now differentiate equation (2.132) with respect to y, substitute y b / 2 in the and equate
y
with zero using boundary condition (2.168)
m b m m b m b m
Am sinh 2a a Bm 2a cosh 2a a
sin m x 0
qa 4
D
m 1
m b m a
Bm sinh
2a a
m b m m b m b m m b m
Am sinh Bm cosh Bm sinh 0
2a a 2a 2a a 2a a
Am sinh m Bmm cosh m Bm sinh m 0 (2.170)
Solving equation (2.169) and (2.170) for the constants Am and Bm and substitute in the equation
(2.132) to obtain the expression for the deflection of the plate.
Exercise
Que. State the assumptions in the small deflection theory of thin plate. [P.U., Ans. Article 2.2.1]
Que. Distinguish between thin and thick plate bending. [P.U., Ans. Article 2.2]
Que. Sketch the free body diagram of a plate element representing lateral loads, moments and
shears. [P.U., Ans. Article 2.2, Figure, Page No. 14]
Que. Write the equations defining equilibrium of the element. [P.U., Ans. Article 2.2, Eq. 2.10,
2.12 and 2.14]
Que. Starting from the first principles derive the governing differential equation in Cartesian co-
ordinates for thin plate under bending.
q
4 w
D
The symbols carry the usual meaning. [P.U., Ans. Derivation of Article 2.2]
Derive the boundary conditions to be used for
1) simply supported edge 2) a clamped edge 3) a free edge[P.U., Ans. Article 2.2.2]
Que. Derive the governing differential equation of thin rectangular plate subjected to transverse
load q x, y per unit area, according to Kirchhoffs theory. Hence explain how Kirchhoffs
reduces three boundary conditions to two per edge. Discuss boundary condition of
cantilever plate. [Dr. B.A.M.U., Ans. Derivation of Article 2.2]
Que. Explain the stepwise procedure in Naviers method for bending analysis of a thin
rectangular plate simply supported at all four edges. [P.U., Ans. Derivation of Article 2.3]
Que. Discuss Naviers solution of simply supported rectangular plate subjected to UDL q x, y .
Hence find the expressions for transverse deflection w x, y , bending and twisting
moments, bending and shear stresses, shear force and reactive forces along the boundaries.
[Dr. B.A.M.U., Ans. Derivation of Article 2.3.1]
Que. Discuss Naviers solution of simply supported rectangular plate subjected to sinusoidal
loading. Hence find the expressions for transverse deflection w x, y , bending and twisting
moments, bending and shear stresses, shear force and reactive forces along the boundaries.
[Dr. B.A.M.U., Ans. Derivation of Article 2.3.2]
Que. Using Naviers method, find for a square plate of side 3 m, thickness 12 cm under uniform
load of 3 KN/m2 values for
a) Maximum deflection in mm
Prof. Atteshamuddin S. Sayyad 58
Theory of Plates and Shells
uniformly distributed load of intensity q0 per unit area. Using Naviers method computes the
deflection and the principal moments at the centre of plate. [P.U., Ans. Problem based on
Derivation of Article 2.3.1]
Que. Using Naviers method, obtain expression for lateral displacement w of the plate a b
subjected to hydrostatic pressure which varies in x direction. Take origin at the centre of left
edge find the central deflection if q0 is the peak intensity of loading on the edge x = a.
[P.U., Ans. Problem based on Derivation of Article 2.3.5]
Que. A thin square plate of size a a and simply supported on all edges carries a patch load.
Using Naviers method computes the deflection and the principal moments at the centre of
plate. [Ans., Derivation of Article 2.3.3]
Que. A thin square plate of size a a and simply supported on all edges carries a point load.
Using Naviers method computes the deflection and the principal moments at the centre of
plate. [Ans., Derivation of Article 2.3.4]
Que. Describe the stepwise procedure in the levys method for thin plate bending analysis.
Illustrate the example where two opposite edges of a rectangular plate are simply supported
[P.U., Ans. Problem based on Derivation of Article 2.4.1]
Que. A square plate a a with all four edges simply supported, carries a uniformly distributed
load of intensity q0. Using levys method, compute the maximum deflection in the plate.
[P.U., Ans. Problem based on Derivation of Article 2.4.1]
Que. Discuss Levys solution of simply supported rectangular plate subjected to UDL q x, y .
Hence find the expressions for transverse deflection w x, y , bending and twisting
moments, bending and shear stresses, shear force and reactive forces along the boundaries.
[Dr. B.A.M.U., Ans. Derivation of Article 2.4.1]
Que. Using Levys method, obtain expression for lateral displacement w of the plate a b
subjected to hydrostatic pressure which varies in x direction. Take origin at the centre of
left edge find the central deflection if q0 is the peak intensity of loading on the edge x = a.
[P.U., Ans. Problem based on Derivation of Article 2.4.2]
Que. Using Levys method, obtain expression for lateral displacement w of the plate a b
subjected to moments at the edges. Take origin at the centre of left edge find the central
deflection. [Ans., Derivation of Article 2.4.3]
Que. Discuss Levys solution of simply supported rectangular plate subjected to UDL q x, y .
Hence find the expressions for transverse deflection w x, y [Ans., Derivation of Article
2.4.4]
Chapter 3
Pure Bending of Plate
direction is iy dw / dy .
Taking now any direction an in the xy plane as shown in figure (b) making an angle with the x
axis. We find that the difference in the deflection of the two adjacent points a and a1 in the an
direction is
w w w x , y n, t
dw dx dy
x y
using chain rule
And slope in the same direction is w w x w y
w w x w y n x n y n
(3.1)
n x n y n
w w w
cos sin (3.2)
n x y
cos sin (3.3)
n x y
Similarly
w w x w y
(3.4)
t x t y t
w w w
cos 90 sin 90
t x y
w w w
sin cos (3.5)
t x y
sin cos (3.6)
t x y
To find out the direction in which the slope of the surface is maximum, we differentiate equation
(3.2) w.r.t. and equate it with zero. Hence the direction of slope can be determine as
w w
sin cos 0
x y
w / y
tan 1 (3.7)
w / x
We can also determine the direction in which the slope of surface would be minimum by
equating equation (3.2) to zero. This direction can be obtained as
w w
cos sin 0
x y
w / x
tan 2 (3.8)
w / y
Therefore from equation (3.7) and (3.8) we note that
tan 1 tan 2 1 (3.9)
Which indicates that the directions of maximum and minimum slopes are orthogonal.
Now to obtain the relations of curvatures,
In considering the curvature of the middle surface in any direction an we obtain
1 2w w
2
rn n n n
Therefore from equation (3.2) and (3.3) we get
2w w w
cos sin cos sin
n 2
x y x y
2w 2w 2w 2w 2w
2 cos 2
sin cos sin cos sin 2
n 2
x yx yx y 2
2w 2w 2w 2w
2 cos 2
2 sin cos sin 2
n 2
x yx y 2
2w 1 1 1
cos 2
2sin cos sin 2
n 2 r rxy ry
x
Prof. Atteshamuddin S. Sayyad 63
Theory of Plates and Shells
1 1 1 1
cos 2 sin 2 sin 2 (3.10)
rn r rxy ry
x
Where
1 2w
2 Bending curvature of thesurfacein plane parallel to xz plane
rx x
1 2w
2 Bending curvature of thesurfacein plane parallel to yz plane
ry y
1 2w
Twisting curvature or twist of thesurface w.r.t. x & y axis.
rxy yx
1 2w
curvature of the middle surfacein any direction an
rn n2
Note: curvatures are the second order derivatives of slope. The curvature is considered positive
if it is convex downward, the minus sign is taken in the equation, since for the deflection
convex downward, as shown in figure, the second derivative 2 w / x 2 is negative.
Similarly, instead of the direction an if we take the direction perpendicular to an, the curvature
in this new direction will be obtained from equation (3.10) by substituting for . Thus
2
we get
1 1 1 1
cos 2 sin 2 sin 2
rt r 2 rxy 2 ry 2
x
1 1 1 1
cos 2 sin 2 sin 2 (3.11)
rt r rxy rx
y
Where
1 2w
curvature of the middle surfacein any direction at
rt t 2
OR
2w w
t 2
t t
2w w w
sin cos sin cos
t 2
x y x y
2w 2w 2w 2w 2w
2 cos 2
sin cos sin cos sin 2
t 2
y yx yx x 2
2w 2w 2w 2w
2 cos 2
2 sin cos sin 2
t 2
y yx x 2
2w 1 1 1
cos 2
2sin cos sin 2
t 2 r rxy rx
y
1 1 1 1
cos 2 sin 2 sin 2
rt r rxy rx
y
Adding equation (3.10) and (3.11) we obtain
1 1 1 1 1 1 1 1
cos 2 sin 2 sin 2 cos 2 sin 2 sin 2
rn rt r rxy ry r rxy rx
x y
1
sin 2 cos 2 sin cos 2
1 1 1
2
rn rt rx ry
1 1 1 1
(3.12)
rn rt rx ry
Which shows that, at any point of the middle surface the sum of the curvature in two
perpendicular directions is independent of the angle and this is usually called as the
average curvature on this surface at a point.
The twist of surface at a with respect to the an and at direction is
1 w
rnt t n
1 w w
sin cos cos sin
rnt x y x y
1 2w 2w 2w
2 sin cos
rnt x xy
cos 2 sin 2 2 cos sin
y
1 2 w sin 2 2w 2 w sin 2
2 cos 2 2
rnt x 2 xy y 2
1 1 sin 2 1 1 sin 2
cos 2
rnt rx 2 rxy ry 2
1 1 1 sin 2 1
cos 2 (3.13)
rnt rx ry 2 rxy
In our further discussion we shall be interested in finding, in terms of the direction in which
the curvature of the surface is a maximum or a minimum and in finding the corresponding values
of the curvature. We obtain the necessary equation for determining by equating the derivative
of equation (3.10) w.r.t. equating with zero which gives
1 1 1
2sin cos 2 cos 2 2sin cos 0
rx rxy ry
1 1 1
sin 2 2 cos 2 sin 2 0
rx rxy ry
1 1 1
sin 2 2 cos 2
rx ry rxy
sin 2 2 1
cos 2 rxy 1 1
rx ry
2
rxy
tan 2 (3.14)
1 1
rx ry
From this equation we find two values of differing by / 2 . Substituting these in equation
(3.10) we find two values of 1/ rn , One representing maximum and one representing minimum
curvature at a point a of the surface. These two curvatures are called as principal curvatures
from the surface and corresponding planes defined by and / 2 are called principal
planes of curvatures.
Note: when the co ordinate planes xz and yz are taken parallel to the principle planes of
1 1
curvature at a point a then twisting curvature or twist 0 i.e. 0
rxy rnt
h /2 h /2
M x dy
h /2
x dz. dy Mx
h /2
x dz
h /2 h /2
M y dx
h /2
y dz. dx My
h /2
y dz
h /2 h /2
M xy dy
h /2
xy dz. dy M xy
h /2
xy dz
h /2 h /2
M yx dx
h /2
yx dz. dx M yx
h /2
yx dz
Using strain displacement relationship and stress-strain relationship we obtain following moment
curvature relationship for isotropic plate.
2w
2
M x 1 0 x
2 w
M y D 1 0 2 (3.15)
y
M xy 0 0 1 w
2
2
2
xy
i.e.
2w 2w 1 1
Mx D 2 Mx D
x y 2 r ry
x
2w 2w 1 1
My D 2 My D
y x 2 r rx
y
2w 1
M xy D 1 M xy D 1
xy rxy
2w
2
r
M r 1 0 1 w 1 2 w
M D 1 0 2 2
(3.16)
M 0 0 1 r r r
r 1 2 w 1 w
2
r r r
3.3 Particular cases in Pure Bending:
Case I) If M x M y M and M xy 0
1 1 1 1
D D
r ry r rx
x y
1 1 1 1
rx rx ry ry
1 1
1 1
rx ry
1 1
(3.17)
rx ry
1
MD 1
rx
1 M
(3.18)
rx D 1
Similarly
1 M
(3.19)
ry D 1
i.e., the plate in this case is bent to a spherical surface the curvature which is given by
equation (3.19).
From equation (3.18)
2w M
x D 1
2
2w M
x2 dx D 1 dx
w M
x C1
x D 1
Integrate w.r.t. x
M x2
w C1 x C2
D 1 2
M x2
w C1 x (3.20)
D 1 2
Similarly from equation (3.19)
2w M
y D 1
2
M y2
w C3 y (3.21)
D 1 2
Therefore adding equation (3.20) and (3.21) we obtain
M x2 M y2
w C1 x C3 y
D 1 2 D 1 2
M x2 M y2
w
C3 y C1 x (3.22)
D 1 2 D 1 2
Where C1 and C3 are constants of integration and define the plane from which deflection are
measured. If this plane taken tangent to the middle surface of plate then C1 C3 0
Therefore equation (3.22) becomes
M x2 M y2
w
D 1 2 D 1 2
w
M
2 D 1
x2 y 2 (3.23 a)
w C x2 y2
w C r2 (3.23 b)
This is the equation of paraboloid of revolution. This gives synclastic surface.
Case II) If M x M y and M xy 0
1 1 1 1
Mx D And M y D
r r r r
x y y x
If M x M y Let M x M1 and M y M2
1 1 1 1
Mx D M1 D (3.24)
r ry r ry
x x
1 1 1 1
And My D M2 D (3.25)
r rx r rx
y y
1
To find value of
r multiply equation (3.24) by and subtract from (3.25)
y
1 1 1 1
M 2 M1 D D
r r ry
y x rx
1 1 1 1
M 2 M 1 D D D 2 D
ry rx rx ry
M 2 M 1 D
1
ry
1 2
1 M 2 M 1
(3.26)
ry D 1 2
Similarly
1 M 1 M 2
(3.27)
rx D 1 2
1
Since M xy 0; 0 (3.28)
rxy
2w M M
1 2
x 2
2
D 1
w M M
1 2
x C1
x D 1
2
In which C1 is the constant of integration, now making integrate again w.r.t. x, we obtained
M M x 2
w 1 2
C1 x C2
D 1 2
2
M M x 2
w 1 2
C x (3.29)
2 1
2
D 1
1
From equation (3.28) of twisting curvature 0 we obtain
rxy
w C5 (3.31)
Adding equations (3.29), (3.30) and (3.31) we get
M M x 2 M M y 2
w 1 2
C1 x 2 1
C y C5
2 3
2 2
D 1 2 D 1
M M x 2 M M y 2
w 1 2
2 1
C1 x C3 y C5 (3.32)
D 1 2 D 1 2
2 2
Where C1 , C3 and C5 are constants of integration and define the plane from which deflection are
measured. If this plane taken tangent to the middle surface of plate then C1 C3 C5 0
Therefore equation (3.32) becomes
M M 2 x 2 M 2 M 1 y 2
w 1 (3.33)
D 1 2
2 D 1 2
2
if M 2 M1
Put in the equation (3.33) we obtain
M M1 x 2 M1 M1 y 2
w 1
D 1 2
2 D 1 2
2
1 y 2 1 x2
w M1
D 1 2
2 D 1 2
2
w
M1
2 D 1
y2 x2 (3.34)
Exercise
Que. Derive the relations between moments, curvatures and deflections in pure bending of plates.
[P.U. Ans., Article 3.2]
Que. Show that curvature 1/ rn in general direction can be obtained with the help
of 1/ rx ,1/ ry and 1/ rxy . [P.U. Ans., Article 3.1., Derivation upto Equation (3.14)]
Que. Prove that the directions of maximum and minimum slopes are orthogonal in pure bent
plate. [Dr. B.A.M.U. & P.U. Ans., Article 3.1., Derivation upto Equation (3.9)]
Que. Shows that, at any point of the middle surface of the bent plate, the sum of the curvature in
two perpendicular directions is independent of the angle . [Dr. B.A.M.U. & P.U. Ans.,
Article 3.1., Derivation upto Equation (3.12)]
Que. Using first principles determine central deflection of square plate subjected to pure couple
M distributed along all edges. Corner points are simply supported. Couple on edges
produces synclastic bending surface. [P.U. Ans., Article 3.3, Case I]
Que. A square plate of 400 mm side and 10 mm thick is simply supported at its four corners. If is
subjected to uniformly distributed pure couple M1 acting on edges parallel to x axis
creating hogging and uniformly distributed couple M2 on edges parallel to y axis creating
sagging. If M1 = M2 = 500 Nm, E = 200 GPa and 0.3 find lateral displacement of
midpoints of edges. [P.U. Ans., Article 3.3, Case II]
Chapter 4
Bending of Circular Plates
Then for the small value of w, the maximum slope of deflected surface at A is given by
w 1 2 w
and the curvature of plate in diametral section rz is 2
r rn r r
Where, = Small angle between the normal to the deflection surface at A and the axis of
symmetry OB.
From the symmetry we conclude that 1/ rn is one of the principle curvatures of the deflected
surface at A and the second principle curvature will be in the section through normal AB and
perpendicular to rz plane. Therefore we conclude that AB is the radius of second principle
curvature which is denoted by rt
1 1 w
(4.1)
rt r r r
Bending moments per unit length along mn is
d 2 w dw d
Mr D 2 D (4.2)
dr r dr dr r
d 2 w 1 dw d
And M D 2 D (4.3)
dr r dr dr r
Where M r and M denotes bending moments per unit length. The moment M r acts along
circumferential section of the plate, such as the section made by conical surface with the apex at
B, and M acts along the diametral section rz of the plate.
Equations (4.2) and (4.3) contains only one variable, or w , which can be determine by
considering equilibrium of an element of the plate such as element abcd as shown in figure cut
out from the plate by two cylindrical sections ab and cd and by two diametral sections ad and bc.
The couple acting on the side cd of the element is
M r r d
The corresponding couple on the side ab is
dM r
Mr dr r dr d
dr
The couples on the sides ad and bc of the element are each M dr and they give a resultant
From the symmetry it can be concluded that the shearing forces that may act on the element must
vanish on diametral sections of the plate but that they are usually present on cylindrical sections
such as sides cd and ab of the element. Denoting by Q the shearing force per unit length of the
cylindrical section of radius r, the total shearing force acting on the sides cd of the element is
dQ
Q r d , and the corresponding force on the side ab is Q dr r dr d . Neglecting
dr
the small difference between the shearing forces on the two opposite sides of the element, we can
state that these forces give a couple in the rz plane equal to Q r d dr
Summing up the moments with proper sign and neglecting the moment due to external
load on the element as a small quantity of higher order, we obtain the following equation of
equilibrium of the element abcd.
dM r
Mr dr r dr d M r r d M dr d Q r d dr 0
dr
dM r
M r r dr d dr r dr d M r r d M dr d Q r d dr 0
dr
dM r dM r
M r r d M r dr d r dr d dr dr d M r r d M dr d Q r d dr 0
dr dr
Since dr and d are very small its higher power is neglected.
dM r
M r dr d r dr d M dr d Q r d dr 0
dr
dM r
Mr r M Q r 0 (a)
dr
Put value of M r and M from equation (4.2) and (4.3) in (a)
d d d d
D r D D Qr 0
dr r dr dr r dr r
d d 2 d 1 d Qr
r 2 r r 2
dr r dr r dr r dr r D
d d 2 d d Qr
r 2
dr r dr dr r dr r D
1 d d 2 Q
2
2 (4.4)
r dr dr r D
dw
Put
dr
1 d dw d2 dw 1 dw Q
dr dr 2 dr r 2 dr D
r dr
d 3w 1 d 2w 1 dw Q
3 2
2
dr r dr r dr D
d 3w 1 d 2w 1 dw Q
3
2
2 (4.5)
dr r dr r dr D
In any particular case of symmetrically loaded circular plate the shearing force Q can easily
being calculated by dividing the load distributed the circle of radius r by 2 r ; then equation
(4.4) or (4.5) can be used to determine the slope and the deflection w of the plate. The
integration of these equations is simplified if we observe that they can be put in the following
form.
d 1 d dw Q
r (4.6)
dr r dr dr D
If Q is represented by a function r, this equation can be integrated without any difficulty in each
particular case. Sometimes it is advantageous to represent the right hand side of equation (4.6) as
a function of intensity q of the load distributed over the plate.
For this purpose we multiply both sides of the equation by 2 r . Then, observing that
r
2 r Q 2 r q dr
0
r
Q q dr
0
Put in the equation (4.6), and multiply both sides of the equation by r, we obtained
d 1 d dw 1
r
r r
dr r dr dr D
0
q r dr
d 1 d 2 w dw 1
r
r r
dr r dr 2
dr D
0
q r dr
Again differentiate left hand side of the above equation with respect to r, we obtained
1 d 3 w d 2 w d 2 w 1 d 2 w dw q r
r r 3 r
dr 2 dr 2 r 2 dr 2
r dr
r dr dr D 0
r
d 3w d 2 w d 2 w 1 dw q
r
dr 3
2
dr 2
dr 2
r dr
D 0
r dr
d 4 w 2 d 3 w 1 d 2 w 1 dw q
4
3
2
3 (4.7)
dr r dr r dr r dr D
This can also be written as
1 d d 1 d dw q
r r (4.8)
r dr dr r dr dr D
This equation can be easily integrated if the intensity of the load q is given as a function of r
4.2 Equation of Deflection for Uniformly Loaded Circular Plate: If a circular plate of radius
a carries a load of intensity q uniformly distributed over the entire surface of the plate,
Now multiply equation (4.8) by r
d d 1 d dw q r
r r
dr dr r dr dr D
Integrate w.r.t r
d 1 d dw q r 2
r r C1
dr r dr dr D 2
where C1 is a constant of integration to be found later from the conditions at the center and at the
edge of the plate. Dividing both sides of above equation by r and making second integration, we
find
1 d dw q r 2
r C1 log r C2
r dr dr D 4
Multiply by r to the both sides of the above equation and performing integration with respect to
r, we obtained
dw q r 4 r2 r2 r2
r C1 log r C2 C3
dr D 16 2 4 2
q r4 r2 r2 r2
w C1 log r C2 C3 log r C4 (4.9)
D 64 4 4 4
Let us now calculate the constants of integration for various particular cases.
Note: since the deflection, moment and transverse shear are to be finite at the centre of
plate r 0 . Therefore the constants C1 and C2 have to remain zero.
d 2w q 3r 2
2 C3 (4.15)
dr 2 D 16
Put equation (4.14) and (4.15) into the equation (4.2)
q 3r 2 q r 3
M r D 2 C3 2 C3 r (4.15 a)
D 16 r D 16
Since this bending moment will vanish at the edges i.e. at r = a, Use equation (4.12b) to obtained
constant C3 by equating equation (4.15 a) with zero
q 3a 2 q a3
D 2 C3 2 C3 a 0
D 16 a D 16
q 3a 2 q a 2
2 C3 2 C3 0
D 16 D 16
q a2
2 C3 1 3 0
D 16
q a 2 3
C3 (4.16)
D 32 1
Put constant C3 from equation (4.16) into the equation (4.13) to obtain constant C4
q a 2 3 2 q a4
a C4 0
D 32 1 D 64
q a 4 3 q a 4
C4
D 32 1 D 64
q a4 2 3
C4 1
D 64 1
q a4 6 2 1
C4
D 64 1
q a4 5
C4 (4.17)
D 64 1
Substitute constants C3 and C4 from equations (4.16) and (4.17) into the equation (4.11) to
obtain deflection of the plate, hence
q a 2 3 2 q a 4 5 q r 4
w r
D 32 1 D 64 1 D 64
q 2 3 2 2 5 a 4 r 4
w ar
64 D 1 1
qa 4 2 3 r 2 5 r4
w
64 D 1 a 2 1 a4
qa 4 r 2 3 5
4 2
r
w (4.18)
64 D a 1 a 1
Deflection is maximum at the center of the plate i.e. at r 0
qa 4 5
w (4.19)
64 D 1
q a 2 3 q 3r 2 q a 2 3 q r 3
Mr D r
D 16 1 D 16 r D 16 1 D 16
qa 4 4 3 12r 2 4 3 4 r 2
Mr 2 4
64 a 1 a 1 a 2 a4
qa 4 r 2 4 1 3
Mr 12 4 4
64 a a2 1
qa 4 3
4
Mr 3 4r 2 4 1 a 2
64 a 1
Prof. Atteshamuddin S. Sayyad 85
Theory of Plates and Shells
q
Mr 3 4r 2 4 a 2 3
64
3 r 2 a2
q
Mr
16
3 a2 r 2
q
Mr (4.20)
16
Use equation (4.3) to find bending moment M , therefore differentiate equation (4.18) w.r.t. r
upto second order
dw qa 4 4r 3 4 3 r
4
dr 64 D a 1 a 2
d 2w qa 4 12r 2 4 3
4 2
dr 2
64 D a a 1
qa 4 12r 2 4 3 1 4r 3 4 3 r
M 4 2 4
64 a a 1 r a 1 a 2
qa 4 4r 2 4 3 12 r 2 4 3
M 2
64 a 4 a 2 1 a4 a 1
qa 4 4 2 a 3 a 2 3
2
M r 3 r
2
64 a 4 1 1
q a2 3
M 1 3 r 1
2
16 1
q
M 1 3 r 2 a 2 3 (4.21)
16
Bending stresses of the plate are to be found out from the following equations
M bh3 h3
r r z I ............. b 1 (4.22)
I 12 12
Substitute equation (4.20) into the above equation (4.22), we get
3 a2 r 2
q z
r
16 I
3 a2 r 2 3 z
q 12
r
16 h
q 3 3 a 2 r 2
3 z
r (4.23)
4 h
Similarly
M bh3 h3
z I ............. b 1 (4.24)
I 12 12
Substitute equation (4.21) into the above equation (4.24), we obtain
q z
1 3 r 2 a 2 3
16 I
q 12
a 2 3 1 3 r 2 3 z
16 h
3 z
q 3 a 2 3 1 3 r 2 (4.25)
4 h
Maximum stress will occur at centre of plate i.e. r = 0, therefore put r = 0 in equation (4.23)
3 z
r max r 0 q 3 a2
4 h3
3 h
r max q 3 3 a2
4 2h
2
3 a
r max q 3
8 h
3 a
r max q 2 3 Aspect Ratio (4.26)
8 h
Similarly, Put r = 0 in equation (4.25) bending stress M
3 z
max r 0 q 3 3 a2
4 h
3 h
max q 3 3 a2
4 2h
2
3 a
max q 3
8 h
3 a
max q 2 3 Aspect Ratio (4.27)
8 h
Therefore maximum stresses in r and are same.
From generalized expression (4.10) for deflection of deflected surface of circular plate subjected
to uniformly distributed load.
q r4
w C1 log r C2 log r . r C3 r C4
2 2
D 64
Note: since the deflection, moment and transverse shear are to be finite at the centre of
plate r 0 . Therefore the constants C1 and C2 have to remain zero.
q a3
2 C3 a 0
D 16
q a2
C3 (4.33)
D 32
Substitute value of C3 in the equation (4.31)
q a2 2 q a4
a C4 0
D 32 D 64
q a2 2 q a4
C4 a
D 32 D 64
q a4
C4 (4.34)
D 64
Substitute C3 and C4 in the equation (4.29)
q a2 2 q a4 q r4
w r
D 32 D 64 D 64
w
q
64 D
a4 2 a2 r 2 r 4
q
a2 r 2
2
w (4.35)
64 D
Deflection is maximum at the center of the plate i.e. r 0
w r 0 max
q a4
64D
(4.36)
This deflection is equal to three-eighths of the deflection of a uniformly loaded strip with built in
ends having flexural rigidity equal to D, a width of unity, and a length of the diameter of the
plate.
Now differentiate equation (4.35) with respect to r upto second order and substitute in the
equations (4.2) and (4.3) to obtain the bending moments M r and M
2 a 2 r 2 2r
dw q
dr 64D
r a2 r 2
dw q
(4.37)
dr 16 D
d 2w
2r 2 a 2 r 2
q
dr 2
16D
d 2w q
2
3 r 2 a 2 (4.38)
dr 16D
Substitute equation (4.37) and (4.38) into the equation (4.2) to obtain the expression for M r
q q
Mr D 3 r 2 a2
r 16 D
r a 2 r 2
16 D
3 r 2 a 2 a 2 r 2
q
Mr
16
q 2
Mr r 3 a 2 1
16
q
Mr a 2 1 r 2 3 (4.39)
16
In the same manner substitute equation (4.37) and (4.38) into the equation (4.3) to obtain M
q 1 q
M D 3 r 2 a 2 r a2 r 2
16D r 16D
q
M 3 r 2 a 2 a 2 r 2
16
q
M 1 3 r 2 1 a 2 (4.40)
16
Substituting r a in the equation (4.39) and (4.40), we find for the bending moments at the
boundaries of the plate
q
Mr ra
a 2 1 a 2 3
16
q
Mr ra
a 2 a 2 3 a 2 a 2
16
Mr
r a max
qa 2
8
(4.41)
And
q
M ra
1 3 a 2 1 a 2
16
q
M ra
a 2 3 a 2 a 2 a 2
16
q a2
M
r a max
8
(4.42)
Similarly
M bh3 h3 h
z I &z ............. b 1 (4.46)
I 12 12 2
Put equation (4.45) in the equation (4.46)
qa 2 12 h
max
8 h3 2
2
3q a
max
4 h
3q 2
max (4.47)
4
From equations (4.45) and (4.47) it is observed that
r max max (4.48)
4.3 Equation of Deflection for Circular Plate subjected to Centre Concentrated Load
dw q r2 r2 r2
C2
2 D 2 4
r log r C1
dr 2
dw q r r r C
log r C1 2
dr 2 D 2 4 2 r
Integrate w.r.t. r
q 1 r2 r2 r2 r2
w log r C C2 log r C3
2 D 2 2
1
4 8 4
qr 2 r2
w log r 1 C1 C2 log r C3 (4.49)
8 D 4
4.3.1 Simply supported circular plate subjected to centre concentrated / point load:
To find the expressions for deflection, moments and corresponding stresses for the
simply supported plate subjected to concentrated load, consider generalized expression (4.49) for
deflection of deflected surface of circular plate subjected to concentrated load.
qr 2 r2
w log r 1 C1 C2 log r C3
8 D 4
Note: since the deflection, moment and transverse shear are to be finite at the centre of
plate r 0 . Therefore the constants C2 have to remain zero.
qr 2 r2
w log r 1 C1 C3 (4.50)
8 D 4
Integration constants C3 and C4 are to be finding out using boundary conditions at the edges of
the plate. In this case deflection and moment must be zero at the edges of plate i.e. at r a
w0 for r a (4.51a)
Mr 0 for r a (4.51b)
Therefore using boundary condition (4.51a), substitute r = a into the equation (4.50) and equate
with zero
qa 2 a2
w r a log a 1 C1 C3
8 D 4
qa 2 a2
log a 1 1 C3 0
C (4.52)
8 D 4
Now, differentiate equation (4.50) with respect to r upto second order to find out moment M r
dw q 21 r
r log r 1 2r C1
dr 8 D r 2
dw qr r
2log r 1 C1 (4.53)
dr 8 D 2
d 2w q 2 C
r 2log r 1 1
dr 2
8 D r 2
d 2w q C
2log r 1 1 (4.54)
dr 2
8 D 2
Put equations (4.53) and (4.54) into the equation (4.2), we get
q C1 qr r
Mr D 8 D 2 log r 1 2 r 8 D 2 log r 1 C1 2
q C
Mr D 2 log r 1 1 1 1
8 D 2
Therefore using boundary condition (4.51b), substitute r = a in the above equation and equate
with zero, we obtain
q C
Mr D 2 log a 1 1 1 1
8 D
r a
2
q C
2log a 1 1 1 1 0
8 D 2
C1 q
1 2log a 1 1
2 8 D
C1
q 1
2 log a (4.55)
4 D 1
Substitute value of constant C1 from equation (4.55) into the equation (4.52) to obtain constant
C3 , we obtain
qa 2 q 1 a 2 C 0
log a 1 2 log a
8 D 4 D 1 4 3
qa 2 1 qa 2 log a 1
C3 2 log a
16 D 1 8 D
qa 2 1
1 log a 1
C3 2log a
8 D 2
1
qa 2 1 1
C3 1 (4.56)
8 D 2 1
qr 2 q 1 r 2 qa 2 1 1 1
w log r 1 2 log a
8 D 4 D 1 4 8 D 2 1
q
2 1 2a 2 1 1 1
w 2r log r 1 r 2log a
2
16 D
1 2 1
2 1 1
q 2
w 2r log r log a 2r 2a r a2
2 2
16 D 1 1
2 r 2 2 1 2 2
w
q
2r log
16 D a 1
a r
2 r 3 2 2
a r
q
w 2r log (4.57)
16 D a 1
Deflection of the plate is maximum at the center of the plate i.e. r = 0
3 2
w r 0 max
q
a
16 D 1
(4.58)
dw q 2 a 1 r 3
2r log 4r 2r
dr 16 D ra a 1
dw q r 3
2r log 4r 2r
dr 16 D a 1
dw q r 3
log 4r 2r 1
dr 16 D a 1
dw q r r
r log (4.59)
dr 4 D a 1
d 2w q a 1 r 1
r log
dr 2
4 D r a a 1
d 2w q a 1 r 1
r log
dr 2
4 D r a a 1
d 2w q r
log (4.60)
dr 2
4 D a 1
Now, to find the bending moments of the plate, Put equation (4.59) and (4.60) into the equation
(4.2) to obtain M r
q r q r r
Mr D log r log
4 D a 1 r 4 D a 1
q r r
Mr D log log
4 D a 1 a 1
q r r
Mr
4 log a log a
q r
Mr
4 log a 1 (4.61)
Similarly put equation (4.59) and (4.60) in the equation (4.3) to obtain M
q r 1 q r r
M D log r log
4 D a 1 r 4 D a 1
q r 1 r r
M D log r log
4 D a 1 r a 1
r r 1
2
q
M log log
4 a 1 a 1
q r 2 1
M 1 log
4 a 1 1
q r 1
2
M 1
log
4 a 1
q r 1 1
M 1 log
4 a 1
q
1 log 1
r
M (4.62)
4 a
Corresponding bending stresses are to be find out using following equations
Mr q r 12 h
r z
4 log a 1
h3 2
I
(4.63)
3q a
r log 1
2h r
2
Similarly
q 12 h
1 log 1 3
M r
z
I 4 a h 2
(4.64)
3q
1 log 1
r
2h
2
a
4.3.2 Fixed / clamped circular plate subjected to centre concentrated / point load:
Consider generalized expression (4.49) for deflection of deflected surface of circular plate
subjected to concentrated load.
qr 2 r2
w log r 1 C1 C2 log r C3
8 D 4
Note: since the deflection, moment and transverse shear are to be finite at the centre of
plate r 0 . Therefore the constants C2 have to remain zero.
qr 2 r2
w log r 1 C1 C3 (4.65)
8 D 4
Integration constants C1 and C3 are to be finding out using boundary conditions at the edges of
the plate. In this case deflection and slope must be zero at the edges of plate i.e. at r a
w0 for r a (4.66a)
dw
0 for r a (4.66b)
dr
Therefore using boundary condition (4.66a), substitute r = a into the equation (4.65) and equate
with zero
qa 2 a2
w r a log a 1 C1 C3
8 D 4
qa 2 a2
log a 1 C1 C3 0 (4.67)
8 D 4
Now, differentiate equation (4.65) with respect to r
dw q 21 r
r log r 1 2r C1
dr 8 D r 2
dw qr r
2log r 1 C1
dr 8 D 2
Now using boundary condition (4.66b), substitute r = a in the above equation and equate with
zero
dw qa a
2log a 1 C1
dr r a 8 D 2
qa a
2log a 1 C1 0
8 D 2
a qa
C1 2log a 1
2 8 D
q
C1 2log a 1 (4.68)
4 D
Substitute value of constant C1 from equation (4.68) into the equation (4.67) to obtain constant
C3 we get
qa 2 q a2
log a 1 2log a 1 C3 0
8 D 4 D 4
q a2 qa 2
C3 2log a 1 log a 1
4 D 4 8 D
qa 2
C3 2log a 1 2 log a 1
16 D
qa 2
C3 (4.69)
16 D
Therefore substitute values of integration constants C1 and C3 into the equation (4.65) to find the
expression for deflection of the plate.
qr 2 q r2 qa 2
w log r 1 2log a 11
8 D 4 D 4 16 D
qr 2 a2
w
1
16 D r 2
2 log r 1 2 log a 1
qr 2 r a2
w 2 log 1 2
16 D
(4.70)
a r
Deflection is maximum at the center of the plate i.e. at r = 0, therefore put r = 0 into the
equation (4.70)
qa 2
wmax (4.71)
16 D
To find the bending moments differentiate equation (4.70) with respect to r upto second order
dw q r a 1
4r log 2r 2 2r
dr 16 D a r a
dw q r
4r log
dr 16 D a
dw qr r
log (4.72)
dr 4 D a
d 2w q a 1 r
r log
dr 2
4 D r a a
d 2w q r
1 log (4.73)
dr 2
4 D a
Therefore substitute equations (4.72) and (4.73) into the equation (4.2) we obtain
q r qr r
Mr D 1 log log
4 D a r 4 D a
q r
Mr 1 1 log (4.74)
4 a
Similarly substitute equations (4.72) and (4.73) into the equation (4.3) we obtain
q r 1 q r
M D 1 log 4r log
4 D a r 16 D a
q r r
M D log log
4 D a a
q r
M 1 log (4.75)
4 a
Corresponding bending stresses are given as follows
Mr q r 12 h
r z r 1 1 log 3
I 4 a h 2
3q r
r 2
1 1 log (4.76)
2 h a
Similarly
M q r 12 h
z 1 log 3
I 4 a h 2
3q r
2
1 1 log (4.77)
2 h a
Equation (4.76) and (4.77) shows that stresses r and become infinite at the centre of plate
r 0 where concentrated load is applied. Therefore the theory which has been presented is
q 3q
M r max r max (4.78)
4 2 h2
q 3 q
M max max (4.79)
4 2 h2
4.4 Problem A circular plate of radius R is clamped along boundary and carries a load whose
r
intensity q / unit area is varies according to the relation q q0 1 obtain an expression for
R
deflection of plate at the radius R and hence calculate maximum deflection.
Solution:
Total load supported by plate at radial distance r
W = Total volume over radius r
= Volume of cylinder + Volume of cone
1
W r2 q r 2 q0 q
3
q 2
W r 2 0 q
3 3
q 2 r
W r 2 0 q0 1
3 3 R
q 2 r
W r 2 0 q0 1
3 3 R
q0 r
W r2 1 2 1
3 R
q0 2r
W r2 3
3 R
q0
W r2 3R 2r
3R
Therefore total load on the plate
2 r Q W
q0
2 r Q r 2 3R 2r
3R
Q
q0
6R
3R r 2r 2 (4.80)
d 1 d dw
r
q0
dr r dr dr 6RD
3R r 2r 2
Integrate w.r.t. r
1 d dw q0 3R r 2 2r 3
r dr dr 6 RD 2 3 C1
r
d dw q0 3R r 3 2r 4
r C1 r
dr dr 6 RD 2 3
Integrate w.r.t. r
dw q0 3R r 4 2r 5 r2
r C1 C2
dr 6 RD 8 15 2
dw q0 3R r 3 2r 4 r C2
C1
dr 6 RD 8 15 2 r
q0 3R r 4 2r 5 r2
w C1 C2 log r C3 (4.81)
6 RD 32 75 4
q0 3R r 4 2r 5 r2
w C1 C3 (4.82)
6 RD 32 75 4
Boundary conditions
dw
w rR 0 and 0 (4.83)
dr r R
q0 3R 5 2 R 5 R2
w rR C1 C3
6 RD 32 75 4
q0 R 4 161 R2
w rR C1 C3
6 D 2400 4
q0 R 4 161 R2
C1 C3 0 (4.84)
6 D 2400 4
Differentiate equation (4.82) w.r.t. r
dw q 3R r 3 2r 4 r
0 C1
dr 6 RD 8 15 2
dw q 3 R 4 2R 4 R
0 C1
dr r R 6 RD 8 15 2
dw q R3 29 R
0 C1
dr r R 6D 120 2
R q0 R 3 29
C1
2 6 D 120
q0 R 2 29 q0 R 2
C1 C1 0.0806 (4.85)
D 360 D
Substitute in the equation (4.84) we get
q0 R 4 161 q0 R 2 29 R 2
C3 0
6 D 2400 D 360 4
q0 R 2 29 R 2 q0 R 4 161
C3
D 360 4 6 D 2400
q0 R 4 29 161
C3
D 1440 14400
q0 R4
C3 0.00896 (4.86)
D
Substitute C1 and C2 in the equation (4.82) of deflection
q0 3R r 4 2r 5 q0 R 2 r 2 q0 R 4
w 0.0806 0.00896
6 RD 32 75 D 4 D
q0 3R r 4 2r 5 q0 R 2 r 2 q0 R 4
w 0.0806 0.00896
6 RD 32 75 D 4 D
Deflection is maximum at r = 0
q0 R4
wmax r 0 0.00896 (4.87)
D
4.5 Problem A circular plate of radius R is simply supported boundary and carries a load
r
whose intensity q / unit area is varies according to the relation q q0 1 obtain an
R
expression for deflection of plate at the radius R and hence calculate maximum deflection.
Solution: Refer previous problem. Solution upto equation (4.82) is same i.e.
q0 3R r 4 2r 5 r2
w C1 C3
6 RD 32 75 4
4.6.1 Bending of a plate by moments M1 and M2 uniformly distributed along inner and
outer boundaries:
Consider bending of a circular plate by moments M1 and M2 uniformly distributed along inner
and outer boundaries. Since plate is subjected to pure bending moments the shearing force
vanishes i.e. Q = 0 in the differential equation (4.6) of deflection.
d 1 d dw Q
r
dr r dr dr D
d 1 d dw
r 0 (4.88)
dr r dr dr
Integrate both sides of above equation with respect to r, we get
1 d dw
r dr r dr C1
Multiply by r to the both sides of above equation and making integration with respect to r again
dw r2
r C1 C2
dr 2
Divide by r to the both sides and integrate with respect to r
r2
w C1 C2 log r C3 (4.89)
4
This is also written as
r2 r
w C1 C2 log C3 (4.90)
4 a
The constants of integration are now to be determined from the conditions at the edges. Since
plate is simply supported along the outer edge, we have
w0 at r a
Mr M2 at r a (4.91)
M r M1 at r b
Since deflection is zero at the supporting edges, using above boundary condition put r = a in the
equation of deflection (4.90) and equate with zero, we get
a2 a
w r a C1 C2 log C3
4 a
a2
C1 C3 0 (4.91 a)
4
Differentiate equation (4.90) twice with respect to r to determine bending moments, hence
dw r C
C1 2 (4.92)
dr 2 r
d 2 w C1 C2
2 (4.93)
dr 2 2 r
Therefore put equation (4.92) and (4.93) in equation (4.2) to find bending moment M r
C C r C
M r D 1 22 C1 2
2 r r 2 r
C C
M r D 1 1 22 1 (4.94)
2 r
This moment must be equal to M 1 for r = b and equal to M 2 for r = a, hence equation (4.94)
becomes
C C
Mr r b
D 1 1 22 1
2 b
C C
M1 D 1 1 22 1 (4.95)
2 b
C C
Mr r a
D 1 1 22 1
2 a
C C
M 2 D 1 1 22 1 (4.96)
2 a
Subtract equation (4.96) from equation (4.95) we get
C C C C
M1 M 2 D 1 1 22 1 D 1 1 22 1
2 b 2 a
1 1
M1 M 2 D C2 1 2 2
b a
M1 M 2 C2
1 D 2 2
1 1
b a
a 2 b2 M1 M 2
C2 (4.97)
1 D a 2 b 2
2 M1 a 2 b2 a 2 M1 M 2
C1
D 1
a 2
b 2
2 a 2 M b2 M
C1 2 1
D 1 a b
2 2
2 b2 M a 2 M
C1 1 2
(4.98)
D 1 a 2 b2
To determine the constant C3 in equation (4.90), the deflection at the edges of plate must be
considered. Therefore put integration constant C1 from equation (4.98) in the equation (4.91 a).
a2
C3 C1
4
a2 a 2 M b2 M
C3 2 1
(4.99)
2 D 1 a b
2 2
Substitute equations (4.97), (4.98) and (4.99) in the equation (4.90) to obtain the expression for
deflection of the plate.
2 r 2 b2 M1 a 2 M 2 a 2 b2 M1 M 2 r
log
D 1 4 a b 1 D a b
2 2 2 2
a
w
a2 a 2 M b2 M
2 1
2 D a b
1 2 2
r a r a M 2 b M1
a 2 b2 M1 M 2
2 2
2 2
w log (4.100)
1 D
a 2
b 2
a 2 D 1
a 2
b 2
Substitute integration constant C1 and C2 from equation (4.97) and (4.98) in the equation (4.94)
for bending moments of the plate.
b2 M a 2 M 2 2
1 2 a b M 1 M 2 1
M r D 1 1 2
2 D 1 a b
2 2
1 D a 2 b 2 r
2
a 2 b2 M M b2 M a 2 M
Mr 1 2 1 2
(4.101)
r a b a b
2 2 2 2 2
Now simplify equations (4.92) and (4.93) by substituting constants of integration C1 , C2 and C3
and then put into the equation (4.3) to obtain bending moment M
C C 1 r C
M D 1 22 C1 2
2 r r 2 r
C C
M D 1 1 22 1 (4.102)
2 r
Put equation (4.97) and (4.98) in the equation (4.102) we get
1 2 b 2 M a 2 M 1 a 2 b 2 M M
M D 1 1 2
1 2
2 D 1 a 2 b 2 r 2 1 D a 2 b 2
b2 M a 2 M a 2 b2 M M
M 1 2
1 2
(4.103)
a b r 2 a 2 b 2
2 2
Corresponding bending stresses are given as follows
Mr M
r z and z
I I
And maximum values of these stresses are given by
M r max M max
r max z and max z
I I
Note: In the previous derivation substitute M2 0 in the equations (4.97), (4.98), (4.99),
2 b2 M
C1 2 12 (4.104)
D 1 a b
a 2 b2 M1
C2 (4.105)
1 D a 2 b2
a2 b2 M
C3 2 1
(4.106)
2 D 1 a b
2
b2 M
a 2 b2 M1 r
r a
1
w 2 1 2 2
log (4.107)
2 D 1 a b 1 D a b a
2 2 2
a 2 b2 M b2 M
Mr 2 2 12 2 1 2 (4.108)
r a b a b
b2 M a 2 b2 M
M 2 12 1
(4.109)
a b r 2 a 2 b 2
Corresponding bending stresses are given as follows
Mr M
r z and z
I I
Now consider the case of bending of a plate by shearing forces Q0 uniformly distributed along
the inner edges as shown in figure. The shearing force per unit length of a circumference of
radius r is
Q0 b P
Q
r 2 r
where P 2 bQ0 denotes the total load applied to the inner boundaries of the plate. Therefore
substitute value of shearing force in the equation (4.6) we get
d 1 d dw P
r
dr r dr dr 2 rD
Integrate both sides of above equation w.r.t. r
1 d dw P
r dr r dr 2 D log r C1
Multiply both sides of above equation by r, and making integration with respect to r we get
dw P r2 r2 r2
C2
2 D 2 4
r log r C1
dr 2
dw P r r r C2
log r C
dr 2 D 2 4
1
2 r
Integrate w.r.t. r
P 1 r2 r2 r2 r2
w log r C1 C2 log r C3
2 D 2 2 4 8 4
P 2 r2
w r log r 1 C1 C2 log r C3 (4.110)
8 D 4
This is also written as
P 2 r r2 r
w r log 1 C1 C2 log C3 (4.111)
8 D a 4 a
The constants of integration will now be calculated from the boundary conditions. Since plate is
simply supported along the outer edge, we have
w0 at r a
Mr 0 at r a (4.112)
Mr 0 at r b
Therefore from equation (4.111)
Pa 2 a2
w r a C1 C3
8 D 4
Pa 2 a2
C1 C3 0 (4.113)
8 D 4
Now differentiate equation (4.111) w.r.t. r upto second order
dw P r r r r C
log C1 2 (4.114)
dr 2 D 2 a 4 2 r
d 2w P 1 r 1 1 C2
log 1 C1 2 (4.115)
dr 2
2 D 2 a 4 2 r
P 1 r 1 1 C2
log 1 C1 2
2 D 2 a 4 2 r
M r D
P r r r r C2
r 2 D 2 log a 4 C1 2 r
P 1 1 r 1 r C1 C2
Mr D log log 1 1
2 D 2 2 a 4 2 a 4 2 r2
P 1 C
1 log 1 1 1 22 1
1 r C
Mr D (4.116)
2 D 4 2 a 2 r
P
1 1 1 22 1
C C
Mr D
8 D
r a
2 a
P
1 1 1 22 1 0
C C
(4.117)
8 D 2 a
Similarly
P 1 C
1 log 1 1 1 22 1
1 b C
Mr D
2 D 4
r b
2 a 2 b
P 1 C
1 log 1 1 1 22 1 0
1 b C
(4.118)
2 D 4 2 a 2 b
Subtract equation (4.118) from (4.117)
P
1 1 1 22 1
C C
8 D 2 a
P 1 C
1 log 1 1 1 22 1 0
1 b C
2 D 4 2 a 2 b
P b C C
log 1 22 1 22 1 0
4 D a a b
P b C C
log 1 22 1 22 1
4 D a a b
P b 1 1
log 1 C2 1 2 2
4 D a a b
b 1 b a
2 2
P
log C2
4 D a 1 b a
2 2
b 1 b a
2 2
P
C2 log (4.119)
4 D a 1 a 2 b2
P C P b b2
1 1 1 log 1 2 0
8 D 2 4 D a a b2
P
1 b 1 b2
C1 log 2
D 1 4
a 2 a b
2
P b 2b2 1
C1 log 2 2 (4.120)
4 D a a b 1
Now from equation (4.113)
Pa 2 a2
C1 C3 0
8 D 4
Pa 2 P b 2b 2 1
a2
log 2 2 C3 0
8 D 4 D a a b 1 4
Pa 2 Pa 2
b 2b2 1
C3 log 2 2
8 D 16 D a a b 1
Pa 2 Pa 2
b 2b2 1
C3 log 2 2
8 D 16 D a a b 1
Pa 2
b b2 1 1
C3 1 log 2 2 (4.121)
8 D
a a b 2 1
Therefore from the equation (4.111) of deflection
P 2 r Pr 2 b 2b 2 1
w r log 1 log 2 2
8 D a 16 D a a b 1
(4.122)
r P b 1 b a
2 2
Pa 2
b b2 1 1
log log 1 log 2 2
a 4 D a 1 a b 8 D
2 2
a a b 2 1
P 2 r 1
Pr 2
Pa 2 1 1
w r log 1 1
8 D a 16 D 1 8 D 2 1
P 2 r r 2 1 a 2 1
w r log r 2
a 2
8 D a 2 1 2 1
2 1 1
2
r
r log a r a r
P
w 2 2 2
8 D
a 2 1
2
r 1 1
r log a r 1
P
w 2 2
(4.123)
8 D
a 2 1
Differentiate above equation twice w.r.t. r and substitute in the equations (4.2) and (4.3) to obtain
bending moments M r and M .
Corresponding bending stresses are given as follows
Mr M
r z and z
I I
as clamped, compute the maximum deflection in the slab. Assume E 0.17 105 MPa and
Poissions ratio = 0.17 for the slab material. [P.U. Article 4.4]
Que. A solid circular slab of concrete with radius R = 3.0 m and uniform thickness of 120mm
r
carries distributed load where intensity q varies according to relation q 4000 1
R
where q is in KN/m at a radial distance r meters from the centre. Assuming the edge of the
slab as simply supported, compute the maximum deflection in the slab. Assume
E 0.15105 MPa and Poissions ratio = 0.17 for the slab material. [P.U. Article 4.5]
Que. A cylindrical R.C.C. water tank with radius 3m stores water to a depth 2m. The bottom slab
is flat with uniform thickness of 180 mm. the slab may be considered to be simply
supported along its edges. Assuming E 0.25105 MPa and Poissions ratio = 0.15, find
the maximum deflection and maximum bending stress. [P.U. Article 4.2.1]
Que. A solid circular slab of concrete with radius 3m and uniform thickness 120mm carries a
uniformly distributed load of 3000 N/m2 assuming the edges of slab as simply supported,
compute the maximum bending moment and deflection in the slab. Starting from
fundamentals, derive the relations you use. For concrete assume E 0.15105 MPa and
Poissons ratio = 0.17. [P.U. Article 4.2.1]
Que. Find the transverse deflection w for the simply supported circular plate with hole of radius a
subjected to moments M1 and M 2 distributed uniformly along inner and outer edges. Hence
find expressions for M r , M , r and and their maximum values. [Article 4.6.1]
Que. Find the transverse deflection w for the simply supported circular plate with hole of radius a
subjected to shearing forces along the inner boundaries. Hence find expressions for
M r , M , r and and their maximum values. [Article 4.6.3]
Chapter 5
General Theory of Cylindrical shell
5.1 Definition
A shell can be defined as a curved structure of which one dimension, the thickness, is small
in comparison with the other two dimensions,
Shell bears the same relation to plates, as curved beams to straight beam
In general shells are termed as curved plates
There are two different classes of shell i.e. Thick shell and Thin shell
A shell will be called thin if the maximum value of ratio h/R can be neglected in
comparison with unity
Where h = thickness of shell
R = Radius of curvature of middle surface
Correspondingly shells will be called thick shell whenever such terms can not be
neglected.
OR h/R > 1/10 = Thick shell
h/R = 1/10 to 1/50 = Thin shell
h/R < 1/50 = Shell is too thin to be used as load carrying member.
5.2 Some important Terms used in shells:
Ruled surface: A ruled surface may be defined as a surface formed by the motion of a straight
line which is known as the generator or ruling.
Singly Ruled surface: A surface is said to be singly ruled if at every point only a single straight
line can be ruled. e.g. Conical shells, conoids and cylinders.
Doubly Ruled surface: A surface is said to be doubly ruled if at every point two straight lines
can be ruled. Hyperbolic paraboloid and hyperboloid of Revolution of one sheet.
Principal curvatures: The curvatures of a point along the direction of maximum and minimum
curvatures are called principal curvatures.
Middle surface: The surface that bisects the thickness of shell or a locus of a point bisecting
thickness of a shell is called middle surface of the shell.
Membrane Action: A shell which carries load entirely by direct stresses lying on its plane is
called as membrane. For membrane action to be possible, shell has to be thin.
Membrane State of Stress: A state of stress in which the stresses in the shell are constant over
its thickness may be defined as Membrane state. A more mathematical approach would be to
regard the membrane theory as a particular case of the more exact bending theory. Thus
membrane theory results if certain effects in the bending theory are ignored.
5.3 Advantages of shell structures:
1. Major load is carried through membrane action and not through bending.
2. Due to this small thickness can be used, requires less material and is economical.
3. Shapes are architecturally beautiful and streamlined.
4. Large floor area uninterrupted by supports is obtained with shells.
5.4 Disadvantage of shell structures:
1. Difficult to analyze.
2. Difficult to construct due to complex geometry.
3. Cannot be used as a floor.
5.5 Classification of shell: Shell surfaces may be broadly classified as singly curved and doubly
curved. Singly curved surfaces are developable. Thus a cylinder can be developed into a plane
rectangle without stretching, shrinking or tearing. Similarly a cone may be developed into a
sector of circle. Doubly curved surfaces are nondevelopable. Hence they will not tend to flatten
out under loads. Further classification of shell surfaces can be attempted on the basis of Gauss
Thin Shell
Synclastic Anticlastic
Gauss Curvature Positive Gauss Curvature Negative
Membrane equation elliptic Membrane equation Hyperbolic
Shells of Shells of
Revolution Translation Shells of Shells of Ruled
e.g. Circular e.g. elliptic Revolution Translation Surface
Domes, paraboloid, paraboloid. e.g. Conical e.g. e.g. Conical &
ellipsoid of Circular Shell Cylindrical Cylindrical
Revolution Paraboloid Shell Shell
2. Straight line normal to undeformed middle surface remains straight and normal to
deformed middle surface.
3. Displacements are small enough so that the changes in geometry of shell are negligible
for equilibrium.
4. The material is linearly elastic, homogenous.
Due to assumption (1) (called Loves Hypothesis) the 3-D problems are converted into 2-D
problems. In assumption (2) effect of shear deformation is neglected. Assumption (3) makes
equations simple and (4) avoid orthotropiy, non-linearity and discontinuities.
5.7 Determination of Stress Resultants:
Consider an infinitesimal small element of shell cut out by sections parallel to x and y axes and
normal to mid plane. Let x and y be the directions of principal curvatures and rx , ry be the radii
of principal curvatures for the element. Let the stresses x , y , xy yx , xz and yz be acting at a
point in the material at distance z from the mid surface as shown in figure.
h /2
z
h /2
z
Qx xz 1 dz Qy yz 1 dz
ry ry
h /2 h /2
h /2
z
h /2
z
M x x z 1 dz M y y z 1 dz (5.1)
ry
h /2 h /2 rx
h /2
z
h /2
z
M xy xy z 1 dz M yx xy z 1 dz
h /2
ry h /2 rx
Rule used in determining the directions of the moments is same as in the case of plates. Since
z z
thickness h of shell is very small in comparison with rx , ry , and may be omitted from
rx ry
z z
Therefore original length of curvature 1 reduced to 1 ' .
rx rx
z z 1 1
Change in length of fibre at a distance z from mid surface = 1 ' 1 z '
rx rx rx rx
z
Original length of fibre = 1
rx
1 1
z '
Strain in x direction = x
change in length
rx rx
orignal length z
1
rx
z 1 1 z 1 1
x ' and y ' (5.2)
z rx rx z ry ry
1 1
rx
ry
In considering stretching of the shell then, let the strain in x and y direction due to stretching is
1 and 2 respectively. For a midplane length dx in x direction, l1 and l2 is the length of fibre
at a distance z from midplane before and after bending.
z z
l1 dx 1 And l2 dx 1 1 1 '
rx rx
z z
dx 1 1 1 ' dx 1
l l2 l1 rx rx
x
l l1 z
dx 1
rx
z z z z z
1 1 1 1 '
1 1 '
1 '
rx rx rx rx rx
x
z z
1 1
rx rx
1 1 1 1 1 1
z ' 1 ' z ' 2 '
1 2 ry ry ry
x
r rx rx
x Similarly y (5.3)
z z z z
1 1 1 1
rx rx ry ry
The thickness of the shell will be always assumed small in comparison with the radii of
curvature. In such a case the quantities like z / rx and z / ry can be neglected in comparison with
unity. We shall neglect also the effect of elongations 1 and 2 on the curvature. Then the
expressions are simplified as,
1 1
x 1 z ' 1 x z (5.4)
rx rx
1 1
y 2 z '
2 y z (5.5)
ry ry
In which 1 and 2 are membrane strains and x and y are bending strains. Where
x and y denotes changes of curvatures. Using these expressions for the components of strain
of a lamina and assuming that there are no normal stresses between laminae z 0 , the
x y z
E
x (5.6)
1 2
z y x 0
E
1 2 (5.7)
z y x
Therefore substitute value of z from equation (5.7) into the equation (5.6) we get
x
E
1 2
x y y x
Substitute value of x and y from equations (5.4) and (5.5) respectively into the above
equation. We get
1 x z 2 y z
E
x
1 2
Note: x y 0 since it is very small
1 2 z x y
E
x (5.8)
1 2
2 1 z y x
E
Similarly y (5.9)
1 2
Substituting these expressions in equation (5.1) and neglecting small quantities z / rx and z / ry in
h /2
z2
Nx
E
1 2
1 2 x y
2 h /2
E
Nx 1h 2 h
1 2
Eh
Nx 1 2 (5.10)
1 2
Eh
Similarly Ny 2 1 (5.11)
1 2
In the same manner bending moments M x and M y can be determined using equations (5.8) and
(5.9).
h /2 h /2
Mx x z dz
E
1 z 2 z z 2 x y dz
1 2
h /2 h /2
h /2
z2 z2 z3
Mx
E
1 2
1 2 x y
2 2 3 h /2
Eh3
Mx
1 2
x y
M x D x y (5.12)
Similarly M y D y x (5.13)
A more general case of deformation of the element is obtained if we assume that, in addition to
normal stresses, shearing stresses also are acting on the lateral sides of the element. Denoting by
the shearing strain in the middle surface of the shell and by xy dx the rotation of the edge BC
relative to Oz about the x axis, we obtained
xy 2 z xy G (5.14)
h /2
z2
N xy G z 2 xy
2 h /2
N xy G h N xy N yx (5.15)
In the same manner twisting moment M xy also be find out from the expression (5.14), we obtain
h /2 h /2
M xy
h /2
xy z dz
h /2
2 z xy G z dz
h /2
z2 z3
M xy G 2 xy
2 3 h /2
G h3 xy
M xy
6
E h3 xy
M xy M yx D 1 xy (5.16)
12 1 2
Thus we can express the resultant forces per unit length N x , N y and N xy and the moments
M x , M y and M xy interms of six quantities of strain. The three components of strain 1 , 2 &
of middle surface of the shell and three quantities x , y & xy representing the changes of
curvature and the twist of the middle surface. In many practical problems bending may be
neglected, as bending if any may be localized in some small portion of shell remaining major
part being subjected to membrane action only.
5.7 General theory of cylindrical shell:
A Circular Cylindrical Shell with Axisymmetric Loadings.
In practical applications we frequently encounter problems in which a circular cylindrical
shell is submitted to the action of forces distributed symmetrically with respect to the axis of the
cylinder. The stress distribution in cylindrical boilers submitted to the action of steam pressure,
stresses in cylindrical containers having a vertical axis and submitted to internal liquid pressure,
and stresses in circular pipes under uniform internal pressure are examples of such problems.
To establish the equations required for the solution of these problems, consider the
various actions on a small element of a shell are as shown in figure. Size of element is dx a d .
Here we have by symmetry, N constant, N x N x 0 and Q 0 also M constant,
N dx d (5.18)
Therefore Fz 0
Qx
Qx a d Qx dx a d N dx d Z dx a d 0
x
Qx
dx a d N dx d Z dx a d 0
x
Qx N
Z 0
x a
Qx N
Z (5.19)
x a
Now taking moment @y i.e. at a distance of dx
M x
M x a d M x dx a d Qx a d dx 0
x
M x
dx a d Qx a d dx 0
x
M x
dx a d Qx a d dx 0
x
M x
Qx 0 (5.20)
x
These are two equations and three unknowns quantities: N , Qx and M x , thus problem in
indeterminate. Therefore two solve the problem consider displacement of a point in the mid
surface of shell. From symmetry we conclude that the component of v of the displacement in
circumferential direction vanishes. We thus have to consider only the components u and w in the
x and z directions. The expressions fir the strain components then become
u w
x and
x a
Therefore using hooks law and integrating over the thickness we get
h /2 h /2
Nx x dz
E
1 2
x dz
h /2 h /2
h /2
E u w
Nx
1 2
h /2
x
dz
a
E u w
z h/2
h /2
Nx 2
1 x a
Eh u w
Nx 2
(5.21)
1 x a
Similarly
h /2 h /2
N dz
E
1 2
x dz
h /2 h /2
Eh u w
N (5.22)
1 2 x a
Since N x 0 (Assumed)
Eh u w
2
0
1 x a
u w
x a
Substitute in the equation (5.22)
Eh 2 w w
N
1 2 a a
Ehw
N (5.23)
a
Considering the bending moments, we conclude from symmetry that there is no bending
d 2w
curvature in circumferential direction, the curvature along x direction is 2 . Therefore from
dx
equation of plate, bending moment in x direction is given by
2w
Mx D 2 (5.24)
x
E h3
Where D = flexural rigidity of the shell
12 1 2
Substitute value of Qx from equation (5.20) into the equation (5.19) we get
2 M x N
Z (5.25)
x a
Put value of M x and N from equation (5.24) and (5.23) into the equation (5.25) to obtained
2 2w E h w
Z
x x 2
D (5.26)
a2
All problems of symmetrical deformation of circular cylindrical shells thus reduced to the
integration of equation (5.26). The simplest application of this equation is obtained when the
thickness of shell is constant. Under such conditions equation (5.26) becomes
4 w E h w
D 4 Z
x a2
4w Z 4 Eh
4 4w 4 (5.27)
x 4
D D a2
N x N x
N x r d N x dx r d N x dx N x d dx X r d dx 0
x
N x N x
dx r d d dx X r d dx 0
x
N x 1 N x
X 0 (5.29)
x r
Similarly, the force in the direction of the tangent to the normal cross section, i.e. in the y
direction gives the corresponding equation of equilibrium, i.e. Fy 0
N N x
N dx N d dx N x r d N x dx r d Y r d dx 0
x
N N x
d dx r d dx Y r d dx 0
x
1 N N x
Y 0 (5.30)
r x
The forces acting in the direction of the normal to the shell, i.e. in the z direction
d d
Total downward force in z direction = N dx sin N dx sin
2 2
d
= 2 N dx = N dx d
2
N dx d Z r d dx 0
N Z r 0
N Z r (5.31)
Note: We find N from equation (5.31) and N x , N x by integration of equation (5.30) and
(5.29)
5.8.1 Example: Horizontal cylinder with closed ends filled with liquid and supported at
ends with simple supports.
Solution:
If P0 be the pressure at the axis of the tube, the pressure at any point is given by,
P z P0 a cos (5.32)
Therefore put value of pressure at any point from equation (5.32) into the equation (5.31) to
find N .
N Z r N Z a
N P0 a cos a
N P0 a cos a
N P0 a a 2 cos (5.33)
Now substitute value of N from equation (5.33) into the equation (5.30) to find out N x .
1 N x
P0 a a 2 cos Y 0
a x
N x
a sin 0 Y 0
x
N x
a sin
x
Therefore integrating above equation w.r.t. x we get
N x
a sin dx
x
N x a sin dx
Nx a sin x C1 (5.34)
Now put value of N x from equation (5.34) into the equation (5.29) to find out N x
N x 1
a sin x C1 X 0
x r
N x 1 1
a cos x C1 0
x a a
N x 1
cos x C1
x a
Integrate w.r.t. x to find out N x
N x 1
cos x C1
x a
x2 x
N x cos C1 C2 (5.35)
2 a
Here integration constant must be determined by using known values of N x and N x i.e.
boundary conditions. For example consider ends horizontally free i.e. N x 0 at x = 0 and x = l
Therefore from equation (5.35)
At x = 0
C2 0 (5.36)
At x = l
l2 l
cos C1 0
2 a
l2 l
cos C1
2 a
al
C1 cos
2
Integrate w.r.t.
al
C
1
2
cos
al
C1 sin C (5.37)
2
al
N x a sin x sin C (5.38)
2
It is seen from equation (5.34) that the constant C represents forces N x uniformly distributed
around the edges of the tube, as is the case when tube is subjected to torsion. If there is no torque
applied, we must take C = 0. Then the above equation becomes
al
N x a sin x sin (5.39)
2
al al
N x x 0
sin And N x x l
sin (5.40)
2 2
By substituting values of C2 and C1 in the equation (5.35) we get
x2 x al
N x cos sin
2 a 2
x2 x a l
N x cos cos
2 a 2
x
Nx l x cos (5.41)
2
And from equation (5.33)
N P0 a a 2 cos (5.42)
Here N x and N x are the shear stresses and bending stresses for the simply supported beam of
hallow circular cross section for the weight of liquid i.e. a 2 per unit length. These are
independent of P0
Consider shell ABCD with semicircular edges AD & BC simply supported on gable walls &
edges AB & CD free. The uniformly distributed load of self weight P acts uniformly over the
middle surface. Area of shell at any point x, we have load components
At or , N 0 which is as required as there are no external forces at straight
2 2
edges, AB and CD. Therefore from equation (5.30)
N x 1 N
Y
x r
Substitute value of N and Y we get,
N x 1
P a cos P sin
x a
N x
P sin P sin
x
N x
2 P sin
x
Integrating both sides with respective x
N x
x
dx 2 P sin dx
Nx 2 P sin x C1 (5.44)
Since N x must be symmetric at the two supported ends and hence constant term C1 must be
N x 2 P sin x (5.45)
It is seen that this solution does not vanish along the edges AB and CD as it should for free
edges. In structural applications, however, the edges are usually reinforced by longitudinal
members string enough to resist the tension produced by shearing force (5.45). Substituting
equation (5.45) in the equation (5.29), we obtained
N x 1 N x
X
x r
Substitute value of N x from equation (5.45) and X to obtain the value of N x , Hence
N x 1
2 P sin x
x a
N x 2P
x cos
x a
Integrate with respective x we get
N x 2P
x
a
x cos dx
2P x2
Nx cos C2 (5.46)
a 2
If the ends of the shell are supported in such a manner that the reactions act in the planes of the
end cross sections, the force N x must vanish at the ends, i.e. N x l /2 N x l /2 0 , Hence we get
2P l2
0 cos C2
a 4
Pl 2
C2 cos (5.47)
4a
Put equation (5.47) in the equation (5.46) to obtained
2P x2 Pl 2
Nx cos cos
a 2 4a
P cos
Nx
4a
4x2 l 2 (5.48)
These expressions N x , N and N x are satisfactory except for the fact that they do not satisfy
the condition that N x must be equal to zero at the free edges. We get some value for N x and to
resist this, a beam or thick strip must be provided at edges AB and CD.
Exercise:
Que. State the advantages and disadvantages of shell structures compared to plates. [P.U. Article
5.3 & 5.4]
Que. Classify thin shell into various types based on shell geometry and curvature. [P.U. & Dr.
B.A.M.U. Article 5.5]
Que. Write the assumptions made in the general theory of thin elastic shells, stating implication
of each. [P.U. Article 5.6]
Que. Derive expressions for the strains x and y at a point due to the bending and membrane
(stretching) action in a shell. Hence obtain expressions for the stress resultants interms of
strain. [P.U. Article 5.7, Derivation of Equation 5.2 & 5.3]
Que. In a thin shell of thickness h, if x , y , xy , xz and yz are the stresses at a point z from
midplane, write expressions for stress resultants at the section, interms of these. [P.U.
Article 5.6, Derivation of Equation 5.6 to 5.16]
Que. sketch cylindrical shell. Show the stress resultants for a small element on the shell surface.
Derive equations of equilibrium for this element. [P.U. Article 5.7]
Que. Derive the governing differential equations for general theory of cylindrical shell,
considering actions on an infinitesimal element. [P.U. Article 5.7]
Que. Write equilibrium equations for membrane analysis of thin cylindrical shells. Using these,
analyse a semicircular cylindrical shell roof of uniform thickness h under a self weight P
per unit area, to obtain membrane stress resultants N x , N and N x in the shell. Assumed
shell has curved edges simply supported and straight edges free. [P.U. Article 5.7.2]
Que. A horizontal cylindrical shell with closed ends is filled with liquid of density and is
simply supported at ends. Derive expressions for stress resultants, N x , N and N x for a
Chapter 6
Bending theory of Cylindrical Shells
actual boundary conditions are realized by applying corrective line loads. But application of such
corrective line load would bend shell and depart from its membrane state. The shell now seek a
new equilibrium and in that process brings into play bending moments, twisting moments and
radial shears. A bending theory is essential to account for these effects.
6.2 Strains in cylindrical shell:
Displacements in x, y and z directions are u, v and w
respectively.
Therefore strain in x direction is given by
u
x (6.1)
x
Similarly strain in y direction is sum of strain
corresponding to plane state of stress and circumferential
strain.
1 v w
(6.2)
a a
1 v
Where strain corresponding to plane state of stress
a
w
And circumferential strain
a
u v
And shear strain x (6.3)
a x
6.3 The Finsterwalder Theory:
Starting from 1932 several rigorous and approximate bending theories have been put
forward for the analysis of reinforced concrete cylindrical shell. The earliest of these was due to
Finsterwalder. By making a few simplifications, finsterwalder was able to develop, for the first
time, a theory that the engineer could use for the analysis of shell roofs. The assumptions
underlying all shell theories and the additional simplifying assumption made by the finsterwalder
theory listed below:
6.3.1 Assumptions made in Finsterwalder theory:
1. Material is homogenous, isotropic and obeys hooks law.
2. Stresses normal to shell surfaces are neglected.
3. All dimensions are very small.
4. A rectilinear element normal to middle surface remains rectilinear after deformation
5. M x M x Qx 0
Note: First four assumptions are common to all bending theories of cylindrical shells. Last
Assumption was introduced by finsterwalder to simplify the problem.
6.3.2 Equations of equilibrium:
It is possible to derive four equations of equilibrium for an element of the unloaded shell acted
upon by the stress resultant shown in figure below.
Prof. Atteshamuddin S. Sayyad 148
Theory of Plates and Shells
N x N x
N x a d N x dx a d N x dx N x a d dx 0
x a
N x N x
dx a d a d dx 0
x a
N x N x
a 0 (6.4)
x
Now summing up the forces in direction, i.e., the direction of the tangent to the shell element
at its midpoint pointing in the direction of increasing and equating them to zero, we get
F 0
N N x
N dx N a d dx N x a d N x dx a d
a x
d Q d
Q dx sin Q a d dx sin 0
2 a 2
Note: First two terms of this equation are the same as those appearing in the membrane
theory. The additional term is the resolved components of shear forces Q in tangential
d N d Q
N dx sin N a d dx sin Q dx Q a d dx 0
2 a 2 a
Neglecting higher powers of dx and d we obtained
d Q
2 N dx sin d dx 0
2
d Q
2 N dx d dx 0
2
On simplifying, we get
Q
N 0 (6.6)
Another equation of equilibrium results from equating the sum of moments of all forces about
the generatrix AD.
M AD 0
M Q
M dx M a d dx Q a d dx a d 0
a a
Neglecting higher powers of dx and d we obtained
M
a Q 0
1 M
Q 0 (6.7)
a
6.3.3 Derivation of Finsterwalder eighth (8th) order differential equation:
Let us introduce a function f which is such that
n x
M f cos
L
n x n a
Or M f cos n (6.8)
a L
From equation (6.8) into the equation of equilibrium (6.7) to obtain value of Q .
1 M
Q
a
1 n x
Q
a f cos a
1 f x
Q cos n (6.9)
a a
Put equation (6.9) into the equation (6.6) to obtain value of N
Q
N
1 f x
N cos n
a a
1 f x
2
N cos n (6.10)
a 2
a
Substitute equation (6.10) into the equation (6.5) to obtain N x
N x 1 1 f n x 1 f n x
2
cos cos
x a a a a 2 a
N x 1 f f n x
3
2 cos
x a
3
a
x
2
a
3 cos
a
dx
1 f f n x a
3
N x 2
sin
a
3
a n
1 f f n x
3
N x sin (6.11)
n a
3
a
Substitute value of N x from equation (6.11) into the equation (6.4) to find out value of N x ,
hence
N x 1 N x
x a
N x 1 1 f f n x
3
sin
x a n a 3 a
N x 1 2 f 4 f n x
sin
x n a 2
2
4
a
N x 1 f f x
2 4
x n a 2 2 4 sin an dx
1 f f n x
2 4
Nx cos (6.12)
n a
2 2
4
a
x
1
Ed
N x N
1
Ed
N N x (6.13)
N x 2 1 N x
x
Gd Ed
Put values of strains from equations (6.1), (6.2) and (6.3) into the equation (6.13), we get
u
1
x E d
N x N
1 v w
1
N N x (6.14)
a a Ed
u v N x 2 1 N x
a x G d Ed
Expressions for u, v and w may now be derived by making use of the stress strain relationship
(6.14). Before we do so, we may set v, the Poissons ratio to zero 0 . We get
u N
x (6.15)
x E d
1 v w N
(6.16)
a a E d
u v 2 N x
(6.17)
a x E d
Now consider equation (6.15) and substitute value of N x from equation (6.12) to obtain value of
displacement u
u 1 1 f f n x
2 4
cos
x E d n a
2 2
4
a
u 1 1 f f x
2 4
x E d n2 a 2 4 cos an
1 f f n x
2 4
u 3
sin (6.18)
E d n 2
4
a
v 2 N x u
x E d a
Substitute values of N x and u from equation (6.11) and (6.18) into the above equation to find
v 2 1 x 1 x
f . f ... sin n f .. f :: sin n
3
x E d n a a a E d n a
Where f . stands for f
v 2 1 x
f . f ... f ... f ::. sin n
x E d n a E d n a
3
a
v 2 1 x
x E d f . f ... f ... f ::. sin n dx
n a E d n a
3
a
2 1 x
v f . f ...
2
f ... f ::. cos n
4
(6.19)
E d n E d n a
Substitute values of N and v from equations (6.10) and (6.19) into the equation (6.20) to obtain
2 1 x a 1 .. n x
w f . f ...
2
f ... f ::. cos n
4 f cos
E d n E d n a Ed a a
2 1 1 .. x
w f .. f ::
2
f :: f :::
4
f cos n (6.21)
E d n E d n Ed a
1 w
v (6.22)
a
Therefore put equations (6.19) and (6.21) into the equation (6.22) we get
2 1
E d 2 f f E d 4 f f
. ... ... ::.
1 n n
cos n x
a 2 1 1 .. a
2
f ..
f ::
4
f ::
f :::
f
E d n E d n Ed
2 1
E d 2 f f E d 4 f f
. ... ... ::.
1 n n
cos n x
a 2 1 1 ... a
2
f ...
f ::.
4
f ::.
f :::.
f
E d n E d n Ed
2 1
2 f f 4 f ... f ::.
. ...
1 n n
cos n x (6.23)
aEd 2 1 ::. f
7 :::.
a
2 f ... f ::. f f ...
n n
4 7
We may next find (change in curvature)
1 2w
2
w (6.24)
a2
Put equation (6.21) in the equation (6.24)
2 1 1 ..
E d 2 f f E d 4 f f E d f
.. :: :: :::
1
2
n n
cos n x
a 22
1 1 .. a
2 f .. f ::
2
f :: f :::
4
f
E d n E d n Ed
2 1 1 ..
E d 2 f f E d 4 f f E d f
.. :: :: :::
1 n n
cos n x (6.25)
2
Eda 2 1 a
2 f f 4 f f f
:: ::: ::: :::: ::
n n
Ed 3 Ed 3
Put in the moment curvature equation M D where D
12 1 2 12
2 1 1 ..
f .. f :: 4
f :: f ::: f
D E d n E d n
2
cos n x (6.26)
Ed
M
E d a2 2 1 a
2 f :: f ::: 4 f ::: f :::: f ::
n n
Put equation (7.26) into above equation (7.8)
2 1 1 ..
E d 2 f f E d 4 f f E d f
.. :: :: :::
2
d
f
n n
2
12 a 2 1
2 f f 4 f f f
:: ::: ::: :::: ::
n n
Rearranging all the terms we get
2 1 1 ..
E d 2 f f E d 4 f f E d f
.. :: :: :::
n n
0 (6.27)
2 f 1
6
a 2
2 f :: f :::
f ::::
f ::
12 f
n 6 n4
d2
n n d
and Qx ignored in the finsterwalder theory is the D-K-J theory in which the three displacements
u. v and w appear in uncoupled form. The theory appears to be due to Donnell, who first used it
in connection with his studies on the stability of thin-walled circular cylinders in 1933-1934.
Karman and Tsien also employed the same theory in 1941 in their investigation on the buckling
of cylindrical shells. Its presentation in a form suitable for the analysis of cylindrical shell roofs
appeared in a book by Jenkins published in 1947. The theory is appropriately called as Donnell
Karman Jenkins theory.
6.4.1 Equations of Equilibrium:
The equilibrium equations already derived in finsterwalder theory will now be modified to take
into account M x , M x and Qx . It will be assumed that M x M x and N x N x
Referring to above figure, the following equations of equilibrium may be set up. Equating all
forces in the x direction to zero we get
Fx 0
N x N x
N x a d N x dx a d N x dx N x a d dx 0
x a
N x N x
dx a d a d dx 0
x a
N x 1 N x
0 (6.29)
x a
Equating the sum of all the forces acting in direction to zero, we get
F 0
N N x
N dx N a d dx N x a d N x dx a d 0
a x
N N x
a d dx dx a d 0
a x
N N x
a 0 (6.30)
x
The third equation of equilibrium is derived by equating all forces in the direction of the inward
normal drawn at the midpoint of the shell element to zero.
d N d Q
N dx sin N a d dx sin Q dx Q a d dx
2 a 2 a
Qx
Qx a d Qx dx a d 0
x
Neglecting higher powers of dx and d we obtained
d Q Qx
2 N dx sin d dx dx a d 0
2 x
Q Qx
N dx d d dx dx a d 0
x
On simplifying, we get
Q Qx
N a 0 (6.31)
x
Another equation of equilibrium results from equating the sum of moments of all forces about
the generatrix. Moments of all forces about the generatrix AD.
M AD 0
M M x
M dx M a d dx M x a d M x dx a d
a x
Q
Q a d dx a d 0
a
Neglecting higher powers of dx and d we obtained
M M x
dx d a dx d a Q dx d 0
x
M M x
a a Q 0 (6.32)
x
Moments of all forces about the generatrix AB.
M AB 0
M x
M x a d M x dx a d M x dx
x
M x Qx
Mx a d dx Qx dx a d dx 0
a x
M x M x
a dx d dx d a Qx dx d 0
x
M x M x
a a Qx 0 (6.33)
x
7.4.2 Flugges simultaneous equations:
Flugge was the first to derive three simultaneous differential equations in u, v and w for most
concrete reinforced shells.
u
From equation (6.15) N x Ed
x
Ed v
From equation (6.16) N w
a
Ed 1 u v
And from equation (6.17) N x a x
2
Substitute N x and N x in the equation (6.29) we get
u 1 Ed 1 u v
Ed 0
x x a 2 a x
2u Ed 1 2u 2 v
0
2 a 2 x
a Ed
x 2
2u 1 1 2u 2 v
0
2 a 2 x
a (6.34)
x 2
Ed v Ed 1 u v
w a 0
a x 2 a x
Ed 2 v w Ed 1 2u 2 v
2 a 2 0
a 2 a x x
2 v w 1 2u 2 v
2
2 a a 0
x 2
(6.35)
2 x
Now from equation (6.32) and (6.33) we obtained
1 M M x
Q
x
(6.36)
a
M x 1 M x
Qx
a
And (6.37)
x
Now using moment curvature relation
2w D 2w D 2w
Mx D , M and M x (6.38)
x 2 a 2 2 a x
Substitute equation (6.38) in the equation (6.36) we get
Prof. Atteshamuddin S. Sayyad 162
Theory of Plates and Shells
1 D 2w D 2w
Q
a a 2 2 x a x
D 3w D 3w
Q
a 3 3 a x 2
D 1 3w 3w
Q 2
a a 3 x 2
(6.39)
3w D 3w
Qx D
x3 a 2 x 2
D 3w 1 3w
Qx
a x3 a x 2
a (6.40)
D 1 3w 3w D 3w 1 3w
N 2 3 2 a a 3 2
a a x
x
a x a x
D 1 4w 4w
D 4w 1 4w
N 2 a a
a a x 2 4 a x a x 2 2
4 4
D 1 4w 4w 2 w
4
N
a a 2 4 x 4
2 a
x 2 4
Ed v
But from equation (6.16) N w
a
Ed v D 1 4w 4w 2 w
4
a a 2 4 x 4
w 2 a
a x 2 4
v Ed 3 1 4 w 4w 2 w
4
12 Ed a 2 4 x 4
w 2 a
x 2 4
v d 2 4w 4w 4 w
4
2
12 a 2 4 x 4
w 2 a a
x 2 4
2
v d 2 2w 2 w
2
2
0
x 2
Or w a (6.41)
12 a
2
Equations (6.34), (6.35) and (6.41) are the simplified version of simultaneous differential
equations due to Flugges in the three displacements u, v and w.
4v 3 w 1 4u 2 v
4
2 2 2
a a 0
x 4
(6.42)
x x 2 x
3
2
Differentiate equation (6.34) w.r.t. operator a
x
4u 1 4u 4v
a3 2
0
2 3 x x 2 2
a a
x 3
1 4u 4v 4u
2
3
2 3 x x 2 2
a a a (6.43)
x 3
2
Now differentiate equation (6.35) w.r.t. operator 2
4 v 3 w 1 4u 4v
4 3
2
0
x 2 2
a a
2 x
3
1 4u 4v 4v 3 w
2
4 3
2 3x x 2 2
a a (6.44)
Therefore from equation (6.43) and (6.44) we get
4u 4v 3 w
a3
(6.45)
x3 4 3
Put equation (6.45) into equation (6.42) we obtain
4v 3 w 1 1 4v 3 w 2 v
4
2 2 2
2
3
a 4
0
x x 2 a x
4
1 4v a 2 4v 4v 3 w 1 3w
2a 2 4 2 x 4 2x 2 x 2 a 2 3
1 4v 4v 4 v
4
3w 1 3w
4 2
x 4 x 2 a 2 3
2 a a
2a 2 2 x 2
2
1 2 2 2 3w 1 3w
x 2 2
a v (6.46)
2a 2 x 2 a 2 3
2 2
2 2 2 v 2 2 2 4 w
4
x 2 2
a
x 2 2
a w a
x 4
2
2 2 2 v 4 w
4
x 2 2
a w a (6.47)
x 4
2
2 2
Now differentiate equation (6.41) w.r.t. operator a 2 2
x 2
2
2 2 2 4 4 4 4
i.e. w.r.t. operator a 2a 2
2
i.e. a
x 2 2 4
we obtained
x x
2 4
2
2 2 2 v
x 2 2 w
a
2
2 2 d 2 2 4w 4 w
4
4w
a 2 2 0
2 12a 2 4
2 a a
x x 2 2 x 4
2 4
2 2 2 v d2 2 2 2
x 2 2
a w 2 x 2 2 w 0
a (6.48)
12a
From equation (6.47) and (6.48) we get
4
4w d2 2 2 2
a 4
x 2 2 w 0
a
x 4 12a 2
4
2 2 2 a6 4 w
a x 2 2 w 12 d 2 x 4
4
2 2 2 a6 4 w
x 2 2
a w 12 0 (6.49)
d 2 x 4
4
2 2 2 a4 4 w
Or a
x 2 2 w 0 (6.50)
k x 4
1 v 1 u v
is that tangential strain w and shear strain x are both very
a a x
u
small as compared to longitudinal strain x
x
Therefore this assumption leads to following relations
1 v v
w 0 w (6.51)
a
1 u v v 1 u
0 (6.52)
a x x a
Note: Schorer theory is applicable to only long shell.
From equation (6.32) we get
1 M
Q
a
1 D 2w D 2w
Q
M
a a 2 2 a 2 2
D 3w
Q (6.53)
a 3 3
From equation (6.31) we get
Q D 3w
N N
a 3 3
D 4w
N (6.54)
a 4 4
2 Nx 1 N x
2
x 2 a x
2 Nx 1 N x
x 2
a x
v 1 u
x a
Differentiate with respective x we get
4v 1 4u
x 4 a x3
4v 1 3u
(6.59)
x 4 a x 3
4w 4v
x 4 x 4
D 8 w 4 w
4 0
E d a 6 8 x
Ed 3 8 w 4 w
4 0 0 (6.61)
12 E d a 6 8 x
8 w a4 4w d2
0 k (6.62)
8 k x 4 12 a 2
N x N yx
N x dy N x dx dy N yx dx N yx dy dx X dx dy 0
x y
N x N yx
dx dy dy dx X dx dy 0
x y
N x N yx
X0 (6.64)
x y
Fy 0
N y N xy
N y dx N y dy dx N xy dy N xy dx dy Y dx dy 0
y x
N y N xy
dy dx dx dy Y dx dy 0
y x
N y N xy
Y 0 (6.65)
y x
For the sake of convenience denote H M xy M yx . Now summing up the moments about
M y M xy
M y dx M y dy dx M xy dy M xy dx dy
y x
Q
Qy y dy dx dy 0
y
M y M xy
dy dx dx dy Qy dx dy 0
y x
M y M xy
Qy 0
y x
M y H
Qy 0 (6.66)
y x
Similarly summing up the moments about edge AB and equate with zero we get
M x H
Qx 0 (6.67)
x y
Now resolving forces in z direction we get
Qx Qy 2 z 2 z 2 z
N x 2 2 N xy Ny 2 z 0
x y x x y y
M x H M y H 2 z 2 z 2 z
N 2 N N z 0
x x y y y x x 2 x y y 2
x xy y
2M x 2 H M y
2
2 z 2 z 2 z
2 N 2 N N z 0 (6.68)
x 2 xy y 2 x 2 x y y 2
x xy y
u u u v
x r w y t w xy 2s w (6.69)
x y y x
Moment curvature relations:
1 2w 1 2w 1 2w
(6.70)
rx x 2 ry y 2 rxy x y
2w 2w
M x D 2 2
x y
2w 2 w
M y D 2 2 (6.72)
y x
2w
M xy H D 1
x y
2 x y 3u 3v 2w 2w
2
r t (6.73)
y 2 x 2 xy 2 yx 2 y 2 x 2
3u 3v 2 x y 2w 2w
2
r t (6.74)
xy 2 yx 2 y 2 x 2 y 2 x 2
Again from equation (6.69)
2 xy 3u 3v 2w
2 s (6.75)
x y y 2x x 2y xy
Therefore put equation (6.74) into the equation (6.75) we get
2 xy 2 x y
2
2w 2w 2w
r t 2 s (6.76)
x y y 2 x 2 y 2 x 2 xy
Let us consider influence of vertical force only i.e. X = Y = 0
2 2 2
Introduce a stress function such that 2 N y ; N and N xy (6.77)
x y x y
2 x
Rearranging the term in the equation (6.76) we get
2 x 2 y 2 xy 2 w 2w 2w
2 r t 2 s 0 (6.78)
y x 2 x y y 2 x 2 xy
Nx Ny 2 N xy
x y xy (6.79)
Ed Ed Ed
2 x 1 2 Nx 2 y 1 Ny
2
2 N xy
2
xy (6.80)
y 2 E d y 2 x 2 E d x 2 E d x y
Put equation (6.77) into the equation (6.80) we get
2 x 1 4 2 y 1 4 2 4
xy (6.81)
y 2 E d y 4 x 2 E d x 4 E d x 2 y 2
Put in the equation (6.78) we obtained
1 4 4 4 2 2 2
2
r t 2 s w 0
E d y 4 x 4 x 2 y 2 y 2 x 2 xy
1 4
k2 w 0
Ed
4 E d 2k w 0 (6.82)
get
D 4 w k2 z (6.83)
All these assumptions will be linearly valid only for long beams. It is observed that beam theory
gives fairly acceptable answers for cylindrical roof shells without edge beams if L/a > 5, and
shells with edge beams if L/a > 3. These limits may be used as guidelines for a designer. The
shell must be uniformly loaded.
Prof. Atteshamuddin S. Sayyad 175
Theory of Plates and Shells
c
1 cos 2 cos sin c sin 2 c
I yy 2 a d 2 d
3
c2
2
0 c
c
1 cos 2 cos sin c sin 2 c
I yy 2 a d 2 d
3
c2
2
0 c
c
sin 2 sin sin c sin 2 c
I yy 2 a d
3
2
2 4 c c2 0
sin 2c 2sin 2 c
I yy a d c
3
2 c
2sin c
I yy a 3 d c sin c cos c (6.85)
c
Equation (6.84) and (6.85) are applicable to only shells of symmetric cross/section.
V .Q
The beam analysis also enables N x to be found out by the use of well known formula for
Ib
the shell subjected to only vertical loading symmetrically distributed over the cross/section.
V .Q.d V .Q.d V .Q
N x (6.86)
I yy b I yy 2d 2 I yy
sin c
Q 2 a 2 d sin (6.87)
c
6.7.4 Arch Analysis: in this second part of analysis, a slice unit width is considered and is
The equilibrium of the arch is maintained by two set of forces, namely the load acting on the
N x
element and (Known as specific shear).
x
This specific shear is resolved into horizontal and vertical components. Sum of vertical
component balance load on shell and sum of horizontal components which are symmetrically
disposed on crown balance themselves.
N x
2 sin d Total load P on the arch of unit width (6.88)
0
x
Exercise:
Que. Derive the 8th order governing differential equations for cylindrical shell subjected to
bending, according to the Finsterwalder theory. [Dr. B.A.M.U. Article 6.3]
Que. Write short note on D-K-J theory for cylindrical shell. [Dr. B.A.M.U. Article 6.4]
Que. Derive equations of VLASOVs bending theory of shallow shells of double curvature from
first principal. [Dr. B.A.M.U. Article 6.6]
Que. Derive Schorer form of differential equations for cylindrical shell. [Dr. B.A.M.U. Article
6.5]
Que. Derive Flugges simultaneous differential equations. [Dr. B.A.M.U. Article 6.4.2]
Que. Describe in brief, the Lundgrens beam theory for thin shells. What are its limitations?
[P.U. & Dr. B.A.M.U. Article 6.7]
Que. Explain the beam method of analysis of cylindrical shell with advantages and limitations of
this method. [Dr. B.A.M.U. Article 6.7]
Chapter 7
Shells of Surface of Revolution
Prof. Atteshamuddin S. Sayyad 178
Theory of Plates and Shells
respectively. Radius of parallel circle is r0 . Let the angle subtended by the element at the centers
of curvature be d and d .
Fig. (7.1)
The dimensions of the element are r1 d in y direction and r0 d in x direction.
In writing the equations of equilibrium of the element, let us begin with the forces in the
direction of the tangent to the meridian. On the upper edge of the element the force N is acting
N r0 d = N r2 sin d (7.3)
N
Similarly corresponding force on the lower edge of the element, is N d acting over the
r
length r0 0 d . Therefore total force on the lower edge is
N r0
N d r0 d (7.4)
From equation (7.3) and (7.4), by neglecting small quantities of second order we find the
resultant in the y direction to be equal to
N r
N r0 d N d r0 0 d d
N r0 N r
N r0 d N r0 d d r0 d N d d d 0 d d
N r0
d r0 d N d d
N r
0 d d (7.5)
The forces acting on the lateral edges of the element are equal to N r1 d and have a resultant
in the direction of the radius of parallel circle equals to N r1 d d . Component of this force
N r
0 d d N r1 cos d d Y r1 r0 d d 0
N r
0 N r1 cos Y r1 r0 0 (7.7)
The second equation of equilibrium is obtained by summing up the projection of the forces in the
z direction. The forces acting on the upper and lower edges of the element have a resultant in z
direction equal to
N r0 d d (7.8)
The forces acting on the lateral sides of the element having the resultant N r1 d d in the
radial direction of the parallel circle gives a component in the z direction of the magnitude
N r1 sin d d (7.9)
The external load acting on the element has in the same direction a component
z r1 r0 d d (7.10)
Therefore sum up the forces in z direction; we obtained the second equation of equilibrium
Therefore from equations (7.8), (7.9) and (7.10) we get
Fz 0
N r0 d d N r1 sin d d z r1 r0 d d 0
N r0 N r1 sin z r1 r0 0 (7.11)
Divide by r1 r0 we get
N r0 N r1 sin zr r
1 0 0
r1 r0 r1 r0 r1 r0
N N sin
z 0
r1 r0
Since r0 r2 sin
N N sin
z 0
r1 r2 sin
N N
z 0 (7.12)
r1 r2
Due to Axisymmetric, equilibrium in x direction is obvious. From above two equilibrium
equations N and N can be found if , r0 , r1 , Y , Z are known.
N N
z
r2 r1
N
N r2 z (7.14)
r1
Substitute equation (7.14) in the equation (7.13) we obtained
N
Y r1 r0 r2 z r1 cos
N r
0
r1
Y r1 r0 z r2 r1 cos r2 N cos
N r
0
Y r1 r0 z r2 r1 cos r2 N cos
N r
0
r2 N cos sin
N r sin Y r
0 1 r2 sin 2 z r2 r1 cos sin
r2 N cos sin
N r sin r
0 1 r2 sin Y sin z cos (7.15)
Which is also written as,
N r2 sin 2 r1 r2 Y sin Z cos sin (7.16)
N r2 sin 2 N r2 2 sin cos r2 sin 2
N
And
N r2 sin cos
N r0 sin N r2 sin cos
N r2 sin sin
N r2 sin cos N r2 cos r2 sin N sin
N r2 2 sin cos r2 sin 2 N
Therefore we can write equation (7.15) as equation (7.16)
Therefore integrating equation (7.16) w.r.t. to find out value of N , we get
N r2 sin 2 d r1 r2 Y sin Z cos sin d
1
r1 r2 Y sin Z cos sin d C
r2 sin 2
N
1
2 r1 r2 Y sin Z cos sin d C
2 r2 sin 2
N (7.17)
The term 2 r1 r2 sin d stands for surface area of an elemental strip of the dome.
Alternatively we may consider equilibrium of the whole cap with total load R (downwards)
Therefore equation of equilibrium along axial direction
2 r0 N sin R 0
R
N (7.18)
2 r0 sin
This equation can be used instead of differential equation (7.7), from which it can be
obtained by integration (7.17). It can be seen that when N is obtained from equation (7.18), the
force N can be calculated from equation (7.12). Hence the problem of membrane stresses can
be readily solved in each particular case.
7.2 Membrane Theory for Surface of Revolution with Unsymmetrical Loading:
Considering again an element cut from a shell by two adjacent meridians and two parallel
circles as shown in figure, in general case not only normal forces N and N but also shearing
forces N N will act on the sides of the element. Area of the element is given in equation
(7.2).
Taking the sum of the projections in the y direction of all forces acting on the element, we must
add to the forces considered in the equations (7.5), (7.6) and External force. The shearing force
N
d r1 d (7.19)
Representing the difference in the shearing forces acting on the lateral sides of the element,
hence by adding equation (7.5), (7.6), External force and equation (7.19), we obtain the equation
N
d r1 d
N r 0 d d N r1 cos d d Y r1 r0 d d 0
N
r1
N r
0 N r1 cos Y r1 r0 0 (7.20)
The second equation of equilibrium is obtained by summing up the projection of the forces in the
x direction; we must include the difference of the shearing forces acting on the top and bottom of
the element as given by the expression
r0 N
N
d d
r0 d d
r0 N d d (7.21)
N
The force r1 d d (7.22)
Due to variation of the force N and the force
N r1 cos d d (7.23)
Due to the small angle cos d between the shearing forces N acting on the lateral sides of
the element. The component in x direction of the external load acting on the element is
X r0 r1 d d (7.24)
Summing up all these forces, we obtain the equation
N
r0 N d d
r1 d d N r1 cos d d X r0 r1 d d 0
N
r0 N
r1 N r1 cos X r0 r1 0 (7.25)
The third equation of equilibrium is obtained by projecting the forces on the z axis. Since the
projection of shearing forces on this axis vanishes, the third equation confirms with equation
(7.11), which was derived for symmetrical bending.
The problem of determining membrane stresses under unsymmetrical loading reduces to
the solution of equations (7.20), (7.25) and (7.11) for given values of components X, Y and Z of
the intensity of the external load.
7.3 Particular cases of Shells in the form of Surface of Revolution:
7.3.1 Example: Spherical dome of constant thickness under its own weight.
Solution: - Consider a spherical dome of constant thickness under its own weight as shown in
figure. Let q be the constant weight of dome per unit area and a be the radius of the sphere.
Let total load R on that part of the spherical dome subtended by angle .
Consider a section at an angle from axis
r0
Therefore sin
a
r0 a sin
The total load R on that part of the spherical dome subtended by angle is given by
R 2 r0 a d q
0
R 2a q sin d
2
R 2 a 2 q cos 0
R 2 a2 q cos cos0
R 2 a2 q 1 cos (7.26)
2 a 2 q 1 cos
N
2 r0 sin
2 a 2 q 1 cos
N r0 a sin
2 a sin 2
a q 1 cos
N
sin 2
a q 1 cos
N
1 cos
2
a q 1 cos
N
1 cos 1 cos
a 2
b 2 a b a b
aq
N (7.27)
1 cos
Now from equation (7.12) we obtained value of N
Where z q cos
N N
q cos 0
a a
Substitute value of N from equation (8.20)
aq N
q cos 0
a 1 cos a
q
N a q cos
1 cos
1
N a q cos (7.28)
1 cos
Now
aq aq
At 0 N (compressive) and N (compressive)
2 2
(Negative sign indicates compressive force and Positive sign indicates tensile force)
Similarly at N a q (compressive) and N a q (Tensile)
2
Which indicates that N is compressive throughout from 0 to 90 but N is
1 1
a q cos 0 cos 0
1 cos 1 cos
cos2 cos 1 0
Solving quadratic equation for cos we get
1 1 4
cos
2
51.8 510 50' (7.29)
In above expression it is assumed that shell gets reaction tangential at the support. If reaction is
not tangential, bending must occur near the support. Generally for shell angles other
than 900 , it is difficult to provide tangential reaction. Hence for such angles usually a ring
beam is used to take horizontal components of the force N and only vertical components are
transferred to the supports. In such cases hoop strain in the beam is differed from strain along
angle in the shell at the beam. Hence there is bending induced locally in a small portion near
the ring beam.
7.3.2 Example: Spherical Tank filled with Liquid.
Solution: Consider a spherical tank with radius a supported along a parallel circle A-A as shown
in figure and filled with liquid of specific weight . Consider hydraulic pressure of the liquid of
specific weight at an angle .
R z a a cos (7.30)
Therefore resultant R of the pressure for the portion of shell defined by an angle is
R a a cos cos 2 a sin a d
0
R 2 a 1 cos cos sin d
3
cos t
Put
sin d dt
R 2 a3 1 t t dt
0
t2 t3
R 2 a
3
2 3 0
cos 2 cos3
R 2 a
3
2 3 0
cos 2 cos3 1 1
R 2 a
3
2 3 2 3
cos 2 cos3 1
R 2 a3
2 3 6
cos 2 2 cos 1
R 2 a3 1
2 3 6
1 cos 2 2 cos
R 2 a3 1
3
(7.32)
6 2
1 cos2 2cos
2 a3 1
R 6 2 3
N
2 r0 sin 2 a sin 2
a 2 1 cos 2 2 cos
N 1
sin 2 6 2 3
a2 1 3 2cos
N 1 3cos 2
sin 6
2
3
a2
N 1 cos 2 3 2 cos
6sin 2
a 2 1 3cos 2 cos
2 3
N
6 sin 2
a2 2cos 2 1 cos
N 1
6
1 cos 2
a2 2 cos 2 1 cos
N 1
6 1 cos 1 cos
a2 2 cos 2
N 1 (7.33)
6 1 cos
Now put value of N from equation (7.33) into the equation (7.12) to determine the value of N
N N
z 0
a a
N
N a z
a
Put value of N and z from equation (7.33) and (7.30) respectively we get
a2 2cos 2
N a 1 a a cos
1 cos
6a
a2 2cos 2
N 1 a 2 1 cos
1 cos
6
a2 2 cos 2
N 1 6 6 cos
6 1 cos
a2 2 cos 2
N 5 6 cos (7.34)
6 1 cos
These expressions for N and N are valid upto 0
For calculating the resultant R for 0 , we must take into account the sum of the vertical
reactions along the ring A-A also in addition to the internal pressure.
a2 2 cos 2
N 5 (7.36)
6 1 cos
a2 2 cos 2
And N 1 6 cos (7.37)
6 1 cos
7.3.3 Example: Conical Shell filled with liquid.
Solution: Let consider a conical shell subjected to load P in the direction of the axis of cone as
shown in figure. Let the semi apex angle of the cone be . Therefore at point where radius of
parallel circle is r0
180
2
(7.38)
2
If the load P is applied in the direction of the axis of cone, then the stress distribution is
symmetrical.
We obtain value of N from equation (8.18)
R P
N
2 r0 sin
2 r0 sin
2
P
N (7.39)
2 r0 cos
Since the curvature of the meridian in the case of a cone is zero, r1 from equation (7.12)
N z r2
z r0
N r0 r2 sin (7.40)
sin
For the above case of loading z = 0 N 0 (7.41)
Now, Let us consider a conical tank to be filled with a liquid of specific weight as shown in
figure. Measuring the distances y from the bottom of the tank and denoting by the total depth of
liquid in the tank, the pressure at any parallel circle mn is
P z
P d y
z d y (7.42)
Also for such tank (7.43)
2
And r0 y tan (7.44)
Where, is the semi-cone angle of the tank. Substituting value of from equation (7.43) in the
equation (7.40) we obtained
z r0
N
sin
z r0
N
sin
2
z r0
N
cos
Put equation (7.44)
d y y tan
N (7.45)
cos
at, y 0 and y d N 0
And at y d / 2 N is maximum
d d
d tan
2 2
N max y d /2
cos
d 2 tan
N max y d /2 (7.46)
4cos
For calculating N we need total load R
1
R r02 d y r02 y
3
1
R y 2 tan 2 d y y 2 tan 2 y r0 y tan
3
y
R y 2 d y tan 2 (8.40)
3
Substitute above equation in the equation to determine N
1 2 y 2
N
2 r0 cos y d y 3 tan
2
y 2 d y tan 2
N 3
2 y tan cos
2
y d y tan
N 3
(7.47)
2 cos
N 0 at y 0
But N 0 at y d
3
N is maximum at y d
4
Prof. Atteshamuddin S. Sayyad 198
Theory of Plates and Shells
3 2 3
d d d tan
4 3 4
N max
y 3 d /4
2 cos
3 d 2 tan
N max y 3d /4
16 cos
(7.48)
Solution: such shells are used as tops/bottom of vertical tanks, or boiler shells. Let a and b be the
semi axes of ellipsoid any angle the principle radius if curvature are given by
a 2b 2 a2
r1 And r2 (7.49)
a2 sin 2 b2 cos2 a2 sin 2 b2 cos2
3/2 3/2
a y b4 x 2 a y b4 x 2
4 2 3/2 4 2 1/2
r1 And r2 (7.50)
a 4b 4 b2
If P is the uniform pressure acting on the shell, then the resultant force R corresponding to an
angle is (Uniform pressure P acting over the area r02 )
P z
R z r02 P (7.51)
From equation (7.18) we get
r02 P
N
2 r0 sin
r22 sin 2 P
N r0 r2 sin
2 r2 sin 2
r2 P
N (7.52)
2
And from equation (7.12)
r P
N r2 2 z
2r1
r22 P
N r2 P z p
2r1
r
N r2 P 1 2 (7.53)
2r1
a2
At 0 from equation (A) r1 r2
b
P a2
N N (7.54)
2b
Similarly
b2
At / 2 r1 and r2 a
a
Pa a2
N And N P a 1 2 (7.55)
2 2b
Therefore N negative in some part near the equator if a 2 2 b2 . Sphere is a spherical case of
Pa
ellipsoid and has a = b from which we have at all points N N for steam pressure P
2
7.3.5 Example: Shell in the form of Torus.
Solution: Rotating a circle @ axis of symmetry we get torus or toroid. Let b be the mean
radius of torus and a be the radius of circular section. Consider equilibrium of a ring shaped
portion formed by rotating portion AB of the circle @ axis of symmetry. Forces N at edge B are
in horizontal plane and being same throughout, these being canceled mutually
(i.e. N cos & N cos ). To find forces N at edge A consider the external load.
Therefore vertical component of external force over the Surface area r02 b 2 is given by
z r02 b 2 P (7.56)
2 r0 N sin (7.57)
2 r0 N sin r02 b 2 P
r2
b2 P r0 b
sin
0
N
2 r0
r0 b a
a
a r0 b r0 b P
N
2 r0 r0 b
a r0 b P
N (7.58)
2 r0
Now from equation (7.12)
N
N r2 z
r1
a r0 b P
N r2 P
2 r0 r1
r2 P
N r0 b 2 r0
2 r0
Pr2
N r0 b r0 r2 sin
2 r2 sin
N
P r0 b (7.59)
2 sin
Note: we can analyze toroids with elliptical section in a similar manner
Exercise:
Que. In terms of coordinates r , and , derive equations of equilibrium for a small element of
axisymmetric thin shell. [P.U. Article 7.1]
Que. For a small element of an axisymmetric shell subjected to axisymmetric loading; derive
equations of equilibrium interms of stress resultants N and N at a point defined by the
Chapter 8
Shells of Double Curvature
8.1 Membrane Theory for Shells of Double curvature other than Surface of Revolution:
z z x, y Eq n of surface of shell
z z 2 z 2 z 2 z
p; q & r ; s ; t
x y x 2 x y y 2
It is expedient to introduce pseudo stress resultants N x , N y and N xy in place of the real stress
resultants N x , N y and N xy . Similarly also introduce pseudo loads X, Y and Z in the directions x, y
and z in place of the real loads Fx , Fy and Fz . The pseudo stress resultant N x is such that it exerts
cos 1 p2
N x dy N x dy Nx Nx
cos 1 q2
1 q2
similarly Ny Ny (8.1)
1 p2
N xy N xy
Fx 1 p 2 q 2 dx dy X dx dy
Therefore,
Fx 0
N x N yx
N x dy N x dx dy N yx dx N yx dy dx X 0
x y
N x N yx
dx dy dy dx X 0
x y
N x N yx
X0 (8.3)
x y
Similarly equating all forces in the y direction to zero we get
N y N xy
Y0 (8.4)
y x
For formulating the equations of equilibrium in the z direction, we have consider element ABCD
1 p 2 z
Vertical component of the normal force acting on AD = N x dy
1 q 2 x
z
= Nx
x
dy
z z
Vertical component of the normal force acting on BC = N x
x
dy Nx
x
dx dy
x
Therefore resultant of the vertical forces on the pair of sides AD and BC
z z z
Nx dy N x dy Nx dx dy
x x x x
z
Nx dx dy (8.5)
x x
z
Similarly, Vertical component of the shear force acting on AD = N xy
y
dy
z z
Vertical component of the shear force acting on BC = N xy
y
dy N xy
x
dx dy
y
Prof. Atteshamuddin S. Sayyad 206
Theory of Plates and Shells
z
N xy dx dy (8.6)
x y
Therefore summing up all the forces acting on AD and BC we get
z z
N x dx dy N xy dx dy
x x x y
z z
Nx N xy (8.7)
x x x y
Similarly summing up all the forces acting on AB and CD we get,
z z
Ny N yx (8.8)
y y y x
The load z contributes a downward force of Z dx dy. summing up all the forces acting on the
element in the z direction we get,
From equation (8.7) and (8.8)
z z z z
Nx N xy Ny N yx z 0
x x x y y y y x
This may be expanded as,
2 z 2 z 2 z 2 z N N yx N y N xy
N x 2 N xy Ny 2 Ny x p q z 0
x x y y x y x y y x
2 z 2 z 2 z N x N yx N y N xy
Nx 2 N N p q z 0
x x y y x y y x
2 xy y 2
Making use of equation () and () we can write
2 z 2 z 2 z
N x 2 2 N xy Ny 2 X p Y q z 0
x x y y
Therefore the equation of equilibrium in the z direction takes the following form
r N x 2 s N xy t N y X p Y q z (8.9)
8.2 Membrane Theory for rectangular hyperbolic paraboloid with straight line generators
and boundaries:
Let, consider equation for a rectangular hyperbolic paraboloid, we may find p, q, r, s and t of
the surface by differentiation. Thus
xy
let z
c
z y z x
p ;q
x c y c
2 z 2 z 1 2 z
& r 0; s ; t 0
x 2 x y c y 2
Put this value in the equation (), we get
2
N xy X p Y q z
C
Denoting right hand side of the equation by L,
CL
N xy L X p Y q z (8.10)
2
Inserting this value in the equation (8.3) and integrate with respective x, we get
N x N yx
x y
X dx f1 y (8.11)
C dL
Nx X dx f1 y (8.12)
2 dy
Similarly making use of equation (8.4) to obtain N y
C dL
Ny Y dy f 2 x (8.13)
2 dx
Consider a rectangular hyperboloid under the action of dead weight g of the shell. The
value of L corresponding to this value is
g
L Z g 1 p2 q2 C 2 x2 y 2
C
Put in the equation (8.10) we obtained
Cg
N xy C 2 x2 y 2
2C
g
N xy C 2 x2 y 2
2
Therefore from equation (8.12)
N x g
C 2 x 2 y 2 dx f1 y
x y 2
g y
dx f1 y
2
Nx (8.14)
C 2 x2 y 2
g x
dy f 2 x
2
Ny (8.15)
C 2 x2 y 2
Nx
gy
2
log x C 2 x 2 y 2 f1 y (8.16)
And Ny
gx
2
log y C 2 x 2 y 2 f 2 x (8.17)
f1 y and f 2 x are to be evaluated from the boundary conditions which depend upon the
Ny 0 at yb (8.19)
These two enables the arbitrary functions f1 y and f 2 x appearing in the equation (8.16) and
(8.17) to be evaluated.
0
gy
2
log a C 2 a 2 y 2 f1 y
f1 y
gy
2
log a C 2 a 2 y 2 (8.20)
Nx
gy
2
log x C 2 x 2 y 2
gy
2
log a C 2 a 2 y 2
gy x C 2 x2 y 2
Nx log (8.21)
2 a C 2 a2 y2
Similarly using equation (8.19) in the equation (8.17)
f2 x
gy
2
log b C 2 x 2 b2 (8.22)
gx y C 2 x2 y 2
Therefore, Ny log (8.23)
2 b C 2 x 2 b2
Equation (), () and (), define the state of stress in the shell. We may note, in passing, that no
boundary conditions were applied along the edges OA and OB which are open boundaries. Along
these edges, one should expect both a normal stress as well as a shear.
Let us consider shallow shell,
For shallow shell p 2 & q 2 are neglected in comparison with unity
z g
Hence from equation (8.10)
gc
N xy
2
Similarly from equation (8.21) and (8.23) Nx 0 & Ny 0
A shallow hyperbolic paraboloid submitted to the action of dead weight develops a state of
pure shear unaccompanished by normal stresses
Let us consider, Hyperbolic paraboloid may visualize as being made up of suspension
cables and arches placed at right angles to each other. At any point on the edge, the normal
components of the tension and thrust of the cable and arch, being equal, cancel each other; the
tangential components add up and give rise to shear. This interpretation also helps us in deciding
on the direction in which the shell is to be reinforced.
Exercise:
Que. Explain how the state of pure shear is developed in s shallow hyperbolic paraboloid under
the action of its dead weight. How is direction of placement of tension steel in such shell
decided? [Dr. B.A.M.U.] [Ans. Article 8.2]
Que. Derive expression for membrane shear for rectangular hyper shell with straight line
generators as boundaries subjected to dead weight g according to membrane theory.
[Dr. B.A.M.U.] [Ans. Article 8.3]
Bibliography
1. Wartikar P. N., and Wartikar J. N., 1998 Engineering Mathematics-II, Pune Vidyarthi
Griha Prakashan., Pune,
Prof. Atteshamuddin S. Sayyad 212
Theory of Plates and Shells